Skip to main content

Full text of "The Interpretation Of Early Music"

See other formats

781.63 D681a 




kansas city 
public library 
kansas city, 

The Interpretation of Early Music 

by the same author 





24 Russell Square 

First published in mcmlxtii 

by Faber and Faber Limited 

24 Russell Square, London WC1 

Second edition mcmlxv 

Printed in Great Britain 

at the Pitman Press, Bath 

All rights reserved 

1963 by Robert Donragton 

In grateful memory 





1. Preview of an 'Early' Style 25 

2. The Element of Passion in Early Music 32 

3. The Element of Serenity in Early Music 34 

4. Reconciling the Passion with the Serenity 35 


1. The Baroque Problem 36 

2. The Factor of Date 37 

3. The Factor of Nationality: Main Currents 38 

4. The Factor of Nationality : Musical Effects 42 

5. The Factor of Purpose Cutting Across both Time 
and Place 45 

6. The Galant Style 46 

7. The Transition to the Classical Period 47 


1 . The Theory of Affects 49 

2. The Factor of Technique 54 

3. Differences of Amateur Standard 55 

4. Differences of Professional Standard 56 

5. The Factor of Taste 57 




L Mode and Key 61 

2. Musica Ficta 'Feigned Music* 62 

3. The Hexachord 63 

4. Unfamiliar Chromatic Signs 65 

5. Unfamiliar Modal Signatures 67 

6. Unfamiliar Key Signatures 67 

7. Written and Unwritten Accidentals 68 


1. Early Accidentals Apply Basically to Single 69 
Notes P<*ge 

2. Effect of the Bar-line on Accidentals 69 

3. Effect of the Hexachord on Accidentals 70 

4. Effect of Repetition on Accidentals 70 

5. Accidentals Occasionally Retrospective 71 

6. Accidentals in Figured Bass 73 


1. Unwritten Accidentals for Necessity and Beauty 74 

2. Correcting Imperfect Intervals 75 

3. The Diminished or Augmented Octave Clash 76 

4. The Picardy Third 77 

5. The Leading-note Principle 79 

6. The Leading-note Itself K() 

7. The 'Double Leading-note' 84 

8. The 'Landini Cadence 1 H4 

9. Sharp Mediants Assumed Next to Sharp Leading- 
notes ^4 

10. Accidentals Suggested by the Performer's Taste 85 

11. Different 'Musica Ficta 1 Shown in Different 
Tablatures of the Same Piece 86 



1. The Performer's Share in the Figuration KK 

2. The Primary Argument about Figuration Itself K c > 

3. The Secondary Argument about Performer's 
Embellishments 91 

4. An Ornamental Figuration to be Taken Lightly 94 

5. Effect of Omitting Obligatory Ornamentation 95 

6. Some Early Terms for Embellishment 96 

7. Solution of Ex. 20 % 


1. Sixteenth-century Ornamentation a Melodic Art 97 

2. Fifteenth and Early Sixteenth Centuries 97 

3. The Later Sixteenth Century 99 

4. Ornamenting Polyphonic Works Today 1 01 



1. Early Baroque Ornamentation Still Mainly 
Melodic page 104 

2. Vocal Ornamentation in Early Baroque Music 104 

3. Instrumental Ornamentation in Early Baroque 
Music 105 


1. Later Baroque Ornamentation Harmonic as well 

as Melodic 109 

2. Vocal Ornamentation in Later Baroque Music 109 

3. Instrumental Ornamentation in Later Baroque 
Music 1 1 1 

4. Concerted Ornamentation in Chamber and Or- 
chestral Music 1 14 

5. National Differences in Embellishment 114 

6. General Advice 115 


1. The Declining Importance of Free Ornamentation 1 17 


1. The Typical Baroque Cadenza an Elaboration of 

a Close 121 

2. The Later Cadenza Part of the Structure of the 
Movement 122 

3. Entire Movements Replaced by Cadenza-like 
Improvisation 122 

4. The Style of Baroque Cadenzas 123 


1. The Proliferation of Ornaments in Baroque Music 125 

2. Some Working Rules for Baroque Ornaments 125 

3. Three Primary Functions Served by Baroque 
Ornaments 130 

4. The Main Baroque Ornaments Grouped by 
Families 131 


L What the Name Appoggiatura Implies 133 

2, The Movement of Appogglaturas 134 



3. All Appoggiaturas take the Beat page 135 

4. All Appoggiaturas are Joined to the Ensuing but 
not to the Previous Note 135 

5. The Early Baroque Appoggiatura 136 

6. The Long Appoggiatura 137 

7. The Function of the Long Appoggiatura Primarily 
Harmonic *41 

8. The Short Appoggiatura 142 

9. Where to Introduce Unwritten Appoggiaturas 145 

10. Appoggiaturas in Recitative 146 

11. The Post-Baroque Appoggiatura 148 


1. What the Name Double Appoggiatura Implies 151 

2. Varieties of Double Appoggiatura 151 

3. The Post-Baroque Double Appoggiatura 152 

XVI. A (iii) : THE SLIDE 

1. What the Name Slide Implies 1 53 

2. Varieties of Slide 153 

3. Post-Baroque Slides 156 


1. What the Name Acciaccatura Implies 157 

2. The Simultaneous Acciaccatura 157 

3. The Passing Acciaccatura 158 

4. The Post-Baroque Acciaccatura 159 


1. What the Name Passing Appoggiatura Implies 161 

2. Evidence for the Passing Appoggiatura 161 

3. Opposition to the Passing Appoggiatura 163 


L What the Names Tremolo and Vibrato Imply 164 

2. The Vocal Tremolo 164 

3. Instrumental Tremolo 166 

4. Vibrato on Instruments 167 

XX, B(iii): THE TRILL 

L What the Name Trill Implies 171 

2. Pre-Baroque and Early Baroque Trills 171 


3. Baroque Trills of the Main Period page 174 

4. The Preparation of Baroque Trills 176 

5. The Speed of Baroque Trills 181 

6. The Termination of Baroque Trills 182 

7. The So-called 'Corelli Clash' Resolved by the 
Conventional Trill 1 8 5 

8. Half-trills 185 

9. Continuous Trills 188 

10. The Ribattuta 190 

1 1 . The Post-Baroque Trill 190 


1. What the Name Mordent Implies 195 

2. Single, Double and Continued Mordents 195 

3. The Inverted Mordent 196 

4. The Position of Mordents 196 

5. Pre-Baroque Mordents 197 

6. Baroque Mordents 197 

7. Diatonic and Chromatic Mordents 198 

8. The Preparation of Mordents 200 

9. The Speed of Mordents 200 

10. The Termination of Mordents 201 

11. The Post-Baroque Mordent 201 


1. Which Ornaments are here Classed as Passing 

Notes 203 

2. The Passing Note Proper 203 

3. The Tirata 203 

4. Which Ornaments are here Classed as Changing 

Notes 204 

5. The Note of Anticipation 204 

6. The Springer 205 

7. The Groppo 205 

8. Post-Baroque Passing and Changing Notes 206 


1. What the Name Turn Implies 207 

2. The Accented Turn 208 

3. The Unaccented Turn 209 

4. Interpretation of Turns 210 




1. In What Respect the Broken Chord is an 
Ornament P a g e 2l2 

2. The Origins of the Broken Chord in Pre-Baroque 
Music 213 

3. The Broken Chord as an Ornament in Baroque 

4. The Broken Chord in Post-Baroque Music 215 

5. The Broken Note Treated as an Ornament 216 

6. Broken Time Treated as an Ornament 217 


1 . What is Meant by the Term Compound Ornament 218 

2. Combinations with the Appoggiatura 218 

3. Combinations with the Trill or Turn 219 



1 . Impromptu Accompaniment 

2. The General Character of Figured Bass 


1. Unfigured Basses 

2. The Signs Used in Figured Bass 

3. The Position of the Figures 

4. Perfect Triads (Common Chords) 

5. Diminished Triads 

6. Augmented Triads 

7. Six-Three Chords (Chords of the Sixth) 

8. Six-Four Chords 

9. Chords of the Augmented Sixth 

10. Chords of the Seventh 

11. Six-Five Chords 

12. Four-Three Chords 

13. Four-Two Chords 

14. 'Passing Notes* in the Bass 


1. What the Figures Show 

2. Departing from the Figures 








3. Departing from the Bass page 246 

4. Accommodating Harmonic Ornaments 249 

5. Adding Accented and Unaccented Passing Notes 254 


L The Importance of Texture in a Good Accom- 
paniment 256 

2. No Accompaniment ('Senza Cembalo', etc,) 256 

3. Unharmonised Accompaniment (Tasto Solo'; 

4 All' unisono') 257 

4. How Thick the Accompaniment Should Be 259 

5. How Strict the Accompaniment Should Be 263 

6. The Filled-in Accompaniment 269 

7. How the Accompaniment Should be Spaced 272 

8. How High and How Low the Accompaniment 
Should Go 273 

9. Texture Further to be Considered in the follow- 
ing Chapter 277 


1 . The Structural Aspect of the Accompaniment 278 

2. How Smooth the Accompaniment Should Be 278 

3. How the Smooth Accompaniment may be 
Broken 280 

4. The Broken Accompaniment in Recitative 283 

5. How Far the Accompaniment Should Double the 
Existing Melody 284 

6. How Far the Accompaniment Should Possess an 
Independent Melodic Interest 286 

7. How Contrapuntal the Accompaniment Should 

Be 289 


1 . Early Baroque Instruments of Accompaniment 293 

2. Supporting the Harmonic Accompaniment with 

a Melodic Bass Instrument 294 

3. Gamba or Cello 296 

4. Adding a Double-bass 297 

5. Other Effects of Sharing Responsibility for the 
Accompaniment 297 

6. The Choice of Harmonic Instrument 298 



1. What Makes a Good Accompanist page 302 

2. Accompanying Recitative 303 

3. Accompaniment as Composition 304 



1. Expression a Spontaneous Impulse 309 


1. Repeats and Omissions within the Performer's 
Option 311 

2. Regular Repeat Signs a Baroque Development 311 

3. Repeats in Dances 313 

4. Repeats in Dance- Form Music 314 

5. Varied Repeats 314 

6. Omitting Movements or Sections 314 



1 . Tempo a Variable Quantity 3 I 6 

2. Early Tempos Only to be Judged by Good 
Musicianship 317 

3. J. S, Bach's Tempos 318 


1. Time-Words Imply Both Mood and Tempo 320 

2. Early Baroque Time- Words 321 

3. Later Baroque Time- Words 322 

4. Metronome Equivalents for Time- Words 324 


L Dance-Titles as Guides to Tempo 326 

2. The Almain (AUemande* etc.) 327 

3. The Ayre 327 

4. The Brawl (Branle, Brando, etc.) 327 



5. The Bourree page 328 

6. The Canaries 328 

7. The Courante (Coranto, etc.) 328 

8. The Chaconne 329 

9. The Entree 330 

10. The Fury 330 

11. The Galliard (Saltarello, Cinq-pas, etc.) 330 

12. The Gavotte 331 

13. The Hornpipe 331 

14. The Jig (Gigue, etc.) 332 

15. The Loure 332 

16. The March 332 

17. The Minuet 332 

18. The Musette 333 

19. The Passacaille (Passacaglia) 334 

20. The Passepied (Paspy) 334 

21. The Pavan 334 

22. The Rigaudon (Rigadoon) 335 

23. The Rondeau (Rondo) 335 

24. The Saraband 335 

25. TheTambourin 336 

26. The Volta 336 

27. Metronome Equivalents for Dance Tempos 337 


1. Origin of Our Time-Signatures in Proportional 
Notation 339 

2. The Fundamental Confusion between Measure 
and Tempo 342 

3. The Subsidiary Confusion over Time-Signatures 

in the Form of Fractions 343 

4. Contradictory Uses of C and <p, etc. 344 

5. Time-Signatures in Triple Time 348 

6. Discrepant Time-Signatures 352 

7. Concurrent Time-Signatures 352 

8. Baroque Time-Signatures in Practice 353 


1. Pulse Not Deliberately Made Audible 354 

2. The Four-time Pulse and the Two-time Pulse 354 

3. The Terms 4 A!la Breve* and *Alla Capdla*, etc. 356 

4. The 4 Alla Breve' in Practice 358 

5. The HeEdola 358 



1. The Need for Expressive Flexibility in Baroque 
Tempo P^e 359 

2. Flexibility of Tempo in Unmeasured and Meas- 
ured Preludes 360 

3. Flexibility of Tempo in the Monodic Style, etc. 360 

4. Flexibility of Tempo in Recitative 360 

5. Flexibility of Tempo in Baroque Music Generally 363 

6. Changes of Tempo between Sections, etc. 363 

7. Variations of Tempo within the Passage 363 

8. Borrowed Time 364 

9. Stolen Time 366 
10. Rallentandos 367 



1. Flexibility the Key to Early Rhythm 369 

2. Rhythms Left Entirely to the Performers 369 

3. Rhythms Left Partly to the Performer 369 

4. Rhythms Loosely Notated for Merc Convenience 370 


1. Methods of Showing Different Note- values in 
Proportional Notation 371 

2. Baroque and Modern Methods of Showing 
Different Note-values 371 

3. Arbitrary Note- values at the Ends of Phrases, etc. 373 

4. A Conventional Mis-notation in Recitative 374 


L The Baroque Dot a Variable Symbol 375 

2. What Baroque Notes arc 'Over-dotted' 375 

3. 'Over-dotting* Itself Variable in Extent 375 

4. Late Baroque Instructions for Dotted Notes 376 

5. Dots as Silence of Articulation 378 

6. General Baroque Preference for the Variable Dot 37 1 ) 

7. Post-Baroque "Over-dotting' 380 


L The Piquant Contrasts of French Overture Style 382 
2. Rests Prolonged before Dotted Figures, etc. * 383 


1. Inequality an Aspect of Rhythmic Freedom page 386 

2. Which Notes Should be Made Unequal 386 

3. The Lilting Rhythm (Lourer) 387 

4. The Snapped Rhythm (Couler) 388 

5. Inequality in Pre-Baroque Music 388 

6. Inequality in Early Baroque Music 389 

7. Inequality in Later Baroque Music 390 

8. How Widely ought Inequality to be Applied? 394 

9. Post-Baroque Inequality 397 


1 . Two Against Three not a Normal Baroque Rhythm 398 

2. Preliminary Notes in Triplet Rhythm 402 

3. The Triplet Convention in the Post-Baroque 
Period 403 



1 , Phrasing must be Audible 404 

2, Phrasing Signs 404 

3, Phrasing Left to the Performer 405 


1. Grouping by Ligatures 407 

2. Words of Articulation 407 

3. Signs of Articulation 407 

4. Articulation Implied by Keyboard Fingering 410 

5. Articulation Implied by Bowings 410 

6. Articulation Left to the Performer 411 

7. Baroque *Ordmary Movement' Somewhat Articu- 
late 414 



L Fitting the Volume to the Music 416 

2. Words of Volume 416 

3. Signs of Volume 417 

4. Range and Flexibility of Early Dynamics 419 



5. Louds and Softs page 419 

6. Crescendos and Diminuendos 421 

7. The 'Messa dl Voce' 422 

8. The Finer Shadings 422 


L Suggestions for Balance shown in Notation 425 

2. Written Indications of Balance 425 

3. Balance Left to the Performer 425 


1. Words Suggesting Accentuation 429 

2. Accentuation Suggested by the Music 429 

3. The Agogic Accent 430 

4. The Weight Accent 430 

5. The Sforzando 430 

6. The Attack Accent 431 



1. Music and its Instruments 435 

2. Substitutions Must be as Suitable as Possible 437 

3. Early Technique Should be Taken Into Account 438 


1. The Effect of Pitch on Performance 439 

2. Actual Pitch and Nominal Pitch 439 

3. Pitches in Late Renaissance and Early Baroque 
Music 441 

4. Pitches in Music of the Main Baroque Period 442 

5. Pitches in Post-Baroque Music 443 

6. Pitch in Practice 445 


L Why Temperament is Necessary 44? 

2. Mean-Tone Temperament 447 

3. Modifications of Mean-Tone Temperament 448 

4. Equal Temperament 448 

5. Temperament in Practice 448 



1. The Italian v Bel Canto 1 page 450 

2. Early Vocal Technique 451 

3. The Trill 454 

4. The Vibrato 456 

5. The Portamento 456 

6. The Necessary Element of Chest Voice 457 

7. The Castrato Voice 457 

8. The Male Alto and the High and Low Counter- 
tenor 458 

9. Underlay of the Words 458 

10. A Declamatory Edge to the Articulation 459 

1 1 . Polyphonic Music also Declamatory 460 

12. The White Voice 460 


1. The Family of Viols 461 

2. The Viols in Consort 462 

3. The Gamba, etc., as Soloist 463 

4. The Ganiba as a Continuo Bass 463 

5. The Cellaraba 463 

6. The Purpose of the Frets 464 

7. The Gamba in the Post-Baroque Period 464 


1 . A Style of Violin-playing for Baroque Music 465 

2. The Fittings of the Baroque Violin 465 

3. The Baroque Violin Bow 466 

4. The Use of Gut Strings 467 

5. Holding the Violin, etc. 468 

6. Holding the Bow 468 

7. Left-hand Technique 469 

8. Right-hand Technique 470 

9. Chords and Polyphony on the Violin, etc. 475 
10. The Double-bass 478 


1. The Lute 479 

2. The Vihuela 481 

3. The Guitar 481 

4. The Cittern 481 

5. The Harp 481 



1 The Changed Character of Many Wind Instru- 
' ments P a S e 482 

2. How Efficient are Early Wind Instruments? 482 

3. Breathing 484 


4. Tonguing ^ 6D 

5. Early Wind Instruments in Practice 485 


1. Embouchure and Fipple 487 

2. Transverse Flutes 487 

3. The Recorders 488 

4. Pipe and Tabor 489 


1. The Influence of Reeds and Bore 491 

2. The Shawms 491 

3. The Oboe 492 

4. The Bassoon 493 

5. The Crumhorn (Crotnorne) 4 93 

6. The Clarinet 494 


1. Brass Cylindrical and Conical, Narrow and Broad 495 

2. The Cornetts 495 

3. The Serpent 496 

4. The Bugles (Tubas, etc.) 497 

5. The Horn 497 

6. The Trombone 498 

7. The Trumpet 499 


1. Military, Dance and Orchestral Percussion 502 

2. The Kettledrum for Orchestral Music 502 

3. The Tabor, etc., for Dancing 502 

4. Turkish Music: The Triangle, Cymbals, etc, 503 


1. The Harpsichord 505 

2. The Clavichord 509 



3. The Piano 

4. The Organ 

5. Fingering 


page 512 

1. Late Renaissance and Early Baroque Ensembles 517 

2. The Standard Baroque Foundation of Strings 518 

3. The Unstandardised Wind 519 

4. The Indispensable Continuo 520 

5. The Conductor 521 

6. Some Typical Baroque Choirs and Orchestras 521 


1. Acknowledgements 525 

2. Conclusion 526 










I have revised this Second Edition to take account of new research and 

If my own opinions have changed at all, it is that I am more than 
ever impressed by the need for inspired performances. What should we 
do at a Shakespeare performance without inspired acting and inspired 
production? What should we get without the declamation, the gestures, 
the timing, the stage business, the costumes, the properties, the lighting? 
No one is shocked or surprised when a Shakespearean producer lays 
firm hands on his text and makes a living drama out of it, 

We have to make living music out of baroque notation. Shakespear- 
ean studies have been going on for a long time, and producers can 
afford to take informed liberties. This book is a contribution to our 
musical studies, mainly of baroque interpretation ; but I hope it may 
also be an encouragement to handle the music with the courage of our 
growing convictions. 

We cannot learn just what inspired fooling Will Kemp first made 
out of Shakespeare's clowns, or with just what grossly rich humanity 
the first Falstaff crossed Shakespeare's boards; but we can build up the 
spirit of the thing out of the poetry and our scholarship combined. 
We cannot learn in quite what manner Corelli "suffered his passions to 
hurry him away so much whilst he was playing on the violin" (see p. 33 
below); but we can discover that he played with passion, and recognise 
it in his music's implications. We can reject the cowardly evasion of 
performing baroque music without any passion at all 

This is not to be less authentic, but more. The most necessary 
authenticity is authenticity of feeling. 

University of Iowa 

Spring, 1965 





The Approach to Early Music 


During the fifteen years over which I was engaged on this book, our 
attitude to early music and its interpretation underwent a remarkable 
transformation. In 1948, the majority of musicians viewed music earlier 
than J. S. Bach with considerable reservations, and remained sceptical 
of any necessity to interpret early music on its own terms, that is to say 
on the instruments for which it was originally composed and in a style 
related to that of its original performers. By 1963, the pleasure which 
sixteenth to eighteenth century music (indeed in some cases much 
older music than this) can give us had been so widely demonstrated, 
together with the practicability and enjoyability of a determined ap- 
proach towards an authentic style, that there could no longer be any 
doubt as to how the matter will end. It will come to be taken as much 
for granted that the best specimens of early music are worth performing 
with all reasonable respect for their composers' intentions as it is already 
taken for granted with regard to the nineteenth century. And why not? 
At the deep levels of human nature from which music speaks, J. S. Bach 
is not so much more different from us than Beethoven, nor is Monteverdi 
so much more different from us than J. S. Bach. 

Those who feel, mistakenly, that early music can be made more acces- 
sible by deliberately modernising it have an undoubted liberty to do so; 
but there seems very little to be said for the inadvertent modernisation 
which results from not having gone sufficiently into the questions of fact 
involved. Questions of interpretation must always revert finally to the 
innate musicianship of the interpreters, but questions of fact can often 
be decided by referring to surviving evidence, and their bearing on 
interpretation can be very intimate. The purpose of this book is to 
contribute to our understanding of early interpretation by assembling 
and discussing some of the surviving evidence. 

There is a place for transcriptions which are a genuine marriage of 
two musical personalities across the generations : the product is a new 
and often interesting work. But in ordinary performance, music of 
whatever generation will sound more effective arid more moving when 
we make every reasonable attempt to present it under its original con- 
ditions of performance. If we want to share in a composer's experience, 



we have to carry out his intentions. If we find his experience somewhat 
strange, we have to remember that it may be more rewarding to come 
to terms with an unfamiliar experience than to recapitulate a familiar 
one in a less telling form. It may be more rewarding to make the most 
of the unfamiliarity than to dilute that unfamiliarity in the doubtful 
hope of adapting it to modern ears. 

Modern ears are no more than ordinary ears in a modern setting. 
Our response to music is fundamentally intuitive, and our human 
faculty of intuition can have changed very little during the longest 
period considered in this book. The fundamentals of interpretation are 
the same for us as they were for Bach or Monteverdi, or indeed for 
Josquin or Perotin le Grand. We not only can but must rely on our 
intuitive response to the expressive implications of early music. But 
this does not mean that we have nothing to learn which the music itself 
cannot teach us. On the contrary, we have a great deal to learn. 

It is unrealistic to think that we can give an adequate rendering of 
any music in the absence of a detailed acquaintance with its relevant 
conventions. Musical notation is a wonderful invention, but it is not 
as wonderful as all that. We need a vast amount of traditional working- 
knowledge in order to bring even the most cunning and thorough of 
these notated marks on paper into living performance. With baroque 
music and earlier, the tradition has been much distorted and partly 
forgotten; the marks may be cunning but we have to be still more 
cunning to know what some of them mean; while notation in general 
was of deliberate intention a great deal less thorough than we are 
nowadays accustomed to finding it. 

To recover something of what direct tradition can no longer convey 
to us, our best and indeed almost our only recourse is to read what 
the actual contemporaries of the music had to say about it. Almost 
exactly half the words in the present book consist of quotations from 
the early authorities and other witnesses themselves, I do most earnestly 
commend these quotations to the careful attention of the reader. It is 
not only that they have an inherent authenticity with which no mere 
summary or discussion of them can compete; it is also that they convey 
so much more than merely factual information. That they are unreliable 
in varying degrees, self-contradictory or contradictory with one another 
in many respects, and tiresomely repetitive in others I shall be the first 
to admit. I almost feel inclined to suggest that this is just why they are 
so valuable. They add up to such a very human picture* They evoke 
the original atmosphere of interpretation more faithfully than any 
systematically consistent account could do. The reality was neither 
systematic nor consistent, and no amount of learned historical after- 
thought can make it so, 



What we are trying to find out is not the exact interpretation such- 
and-such a composer intended for such-and-such a passage. There is 
no such thing as an exact interpretation. No one, not even the composer, 
plays a passage in exactly the same way twice running. It would be a 
much duller world if music would not tolerate more than one interpre- 
tation; we should miss the intensely personal contribution which 
different performers have to offer, or even the same performer when in 
a different mood. What we are trying to find out is partly the kind of 
detail which did not and should not depend on mood at all, and partly 
the outside limits within which the performer's mood can suitably 
operate. We want him to illuminate different aspects of the music in 
accordance with whatever is individual in his response to it; we do not 
want him to impose his individuality on the much greater element in 
the music which is common to us all, and to which there is fundamentally 
only one way of responding. We do not want him to go outside the 

In the case of early music, this is tantamount to saying that we need 
to know, not exactly how the composer would have taken his music, 
but broadly how any good performer of his day might have taken it. 
That leaves a very wide field for personal taste and individuality; it 
even opens the way to performances which I personally, or the reader, 
or any other given person might happen to dislike. But it does set a 
boundary to positive incongruities of style, because while a bad con- 
temporary performer might all too easily have committed them (and 
indeed there is plenty of evidence for this at all times and places), a 
good contemporary performer could not have done so, unless, of 
course, the style in question was outside his experience even though 
contemporary in time. 

Every piece of music carries implications with regard to performance 
which can be differently interpreted: but not beyond certain limits, 
because so soon as those limits are overstepped we feel a contradiction 
between the style of the music and the style of the interpretation. We 
may not, it is true, feel this contradiction explicitly; we may feel vaguely 
uneasy without knowing quite what is wrong or what to do about it; 
or we may notice nothing wrong, and merely fail to be moved by the 
music as in a more understanding performance it could be moving us. 
But whether we are directly aware of it or not, there will still be 
something missing which need not be missing, 

A performer contemporary with the music had opportunities for 
becoming familiar with the style which we are denied at this distance 
of time. A single satisfactory sound-recording might tell us more about 
baroque methods of performance than all our painstaking researches; 
but we have not got that recording, and must do the best we can 



without it. The best we can do is to start from such contemporary 
evidence as does survive. 

This does not imply idealising the contemporary performers, who so 
far as we can make out must have been much worse than ourselves in 
some branches of the art, much better in other branches, and perhaps 
as an overall average very much the same. It is simply that whatever 
else may or may not have been wrong with a contemporary perform- 
ance, under ordinary circumstances it would at least have been within 
the style, whereas under ordinary circumstances a modern performance, 
whatever its other merits, is likely to be partly at odds with the style. 
It is only fair to add that extraordinary circumstances, in which some 
specialised knowledge has been successfully brought to bear, are 
becoming so much more frequent nowadays that we may in a few 
years time be able to regard this favourable situation as no longer the 
exception but the rule. I wrote these words in 1963. 

There were in 1963, and to some extent always will be, limits to the 
authenticity we can hope to achieve. Anyone specialising in the inter- 
pretation of early music needs not only a sufficiently scholarly grasp 
but sufficient competence and experience as a practising musician. He 
must be able to get inside the problems as they actually come up in 
rehearsal; he must be able to envisage solutions which can come off 
in performance. His scholarship can only be helpful if it is used music- 
ally ; yet at the same time it can only be used musically if he has plenty 
of it. He must be in a position to weigh one piece of evidence against 
another. An isolated statement, out of context and perhaps untypical, 
can lead to devastatingly unmusical results; and that after all is the last 
thing we want our scholarship to end in. We are trying to be authentic 
not because there is anything sacrosanct in historical reproduction, but 
because our best chance of matching the interpretation to the music 
lies in matching it to the original intentions. We are trying to be better 
scholars in order to make better music. 

Ultimately it is our personal responsibility as performing musicians 
to make historical authenticity a living thing. It seems unlikely that we 
shall ever make so close a match that it would deceive any visiting 
seventeenth-century or eighteenth-century ghost into thinking that he 
was listening to a performance of his own day; but at least we may hope 
that he would recognise the general style and enjoy the music. We 
may hope to be, not ideally, but reasonably authentic. 

As a corollary to being as authentic as we reasonably can, I think 
we should accept the fact that under modem conditions of performance 
some aspects of authenticity are more important than others, and that 
it is worth letting the less important aspects go by the board provided 
we can get the more important aspects established. It is, for example, 



worth a great deal to get really good performers who are willing to 
co-operate to the best of their ability, but who may have neither the 
time nor the temperament to undergo a prolonged training In early 
style. On the other hand, performers who are not willing to co-operate 
to the best of their ability with regard to the essential points of style 
are unsuitable for the purpose. 

Some modern musicians, again, seem to have a greater natural 
affinity with the stylistic requirements of early music than others. In 
any more or less permanent chamber ensemble or other organisation 
which wants to give special attention to early music, it Is wisest to 
choose members with as much of this natural affinity as possible, and 
then to concentrate and go on concentrating on essentials, with as 
many inessentials added as the circumstances of the case permit. The 
same considerations apply to performers brought together only for a 
particular performance or series of performances, except that the 
necessity to concentrate on essentials is then still more obviously 

The question as to what are the essentials can only be answered in 
concrete detail, as this book proceeds. But it is perhaps worth asking 
here whether there can in principle be any such thing as might be 
described as an 'early' style of interpretation. If we press that question 
too hard, it is plain that the answer must be 'no' ; but in very loose 
terms, it may be possible to give a more positive reply. The contem- 
porary quotations of which this book so largely consists have the effect 
of building up cumulatively a picture of our predecessors in their 
music-rooms and auditoriums, not as stiff historical figures but as very 
human beings with all our own human diversity of tastes and abilities. 
Behind all this diversity, however, we see also what is still more illumi- 
nating: a certain common denominator of tacit assumptions and 
habitual attitudes which may give us our first and most general 
indications of such an 'early' style of interpretation. 

I have fallen back on this non-committal word 'early' whenever I have 
wanted to leave my chronological boundaries deliberately undefined. I 
am mainly though not by any means exclusively concerned in this book 
with baroque music, i.e. approximately from Monteverdi to J. S. Bach. 
There is diversity enough here on any showing; but we find throughout 
a general disposition to join the composer and the performer in a more 
equal partnership than our present custom is. Even when the two were 
not, as they so frequently were, one and the same person, the performer 
was expected to make the music his own with much less respect for the 
written text and much more reliance on spontaneous expression and 
improvisation than we should normally expect now except in dance 
music. It was not that composers were lazier or performers were 



prouder, but that both parties set an overriding value on the freshness 
and immediacy which are among the most positive results of such an 
attitude. They valued spontaneity. 

To recapture this sense of spontaneity is the most important single 
factor in our search for an adequately authentic rendering. Trills and 
appoggiaturas ; conventions of rhythm, tempo and dynamics; small 
forces and original instrumentation: such matters are valuable con- 
tributions to the style; but they are not the style. The whole is greater 
than the parts. 

The style most widely appropriate to baroque music is less massive 
but more incisive than that in which my generation grew up. It is vivid 
yet relaxed; glowing yet transparent. It sparkles and it dances, alive 
with natural ease and unforced conviction. It charms like a smile, and 
it cuts like a knife. The less we inflate it the stronger it sounds. Even 
the best baroque music can be made ponderous by overweighting it, 
but its true nature is as volatile as its performers can possibly conceive 
of it in their most impulsive moods. 

Contrary to some modern opinion, there is nothing unimpulsive, and 
nothing dry, about an authentic rendering of early music. That such 
an opinion should have arisen was understandable and valuable earlier 
in the twentieth century when the most pressing necessity was to escape 
from the incongruous influence of post-Wagnerian weight, sonority 
and smoothness. But this escape has now virtually been accomplished, 
and our present danger is not too little austerity but too much. We 
are in some danger of depriving early music of the sheer animal vitality 
which carries all genuine musical performance along as nature always 
needs to carry along and underwrite the achievements of culture. 

There is a magnificence of storm and stress which is part of Wagner's 
musical language. But there is also a poise and a crispness and a 
crystalline translucency shared by composers as unlike in other respects 
as Monteverdi and Vivaldi, as Purcell and Couperin, as Bach and 
Handel. This baroque brand of eloquence is not less impassioned than 
the romantic variety. To match up to the baroque intentions in our 
interpretation does not, as is sometimes mistakenly suggested, mean 
renouncing all our warmest feelings and all our richest colourings of 
tone. It does mean applying them appropriately to the matter in hand, 
keeping the style sharply etched and the mood unaffected and direct. 
It does mean reconciling the complementary requirements of passion 
and of serenity. 

Ignoring the passion in early music is a mere escape into fantasies 
of unbroken serenity in some past golden age: fantasies just as com- 
fortable and illusory as our by now untenable and discarded Victorian 
fantasies of unbroken progress. There has never been an age of 



unqualified serenity or unqualified passion; these are two extremes of 
human experience which the art of living consists in more or less 
successfully reconciling. And the art of music, like the plastic and 
literary arts, very largely consists in showing how they can be reconciled. 

It is this reconciliation of opposites achieved by the composer in Ms 
music which it is so necessary for our interpretation to carry faithfully 
into effect. All great music achieves it, but not all in the same way. 
The differences do not lie simply along chronological divisions. It is 
personality rather than period which separates the other-worldliness of 
Palestrina from the earthy immediacy of Monteverdi, the rich homeli- 
ness of Haydn from the profound lucidity of Mozart, the lyricism of 
Schubert from the indomitability of Beethoven. There are extraverts 
like Vivaldi or Handel and introverts like Purcell or J. S. Bach in any 
age. Brahms and Wagner and Debussy are all of the nineteenth century; 
Stravinsky and Schoenberg and Vaughan- Williams and Bartok are all 
of the twentieth. Their solutions to the equation are various in the 
extreme; but not the elements in the equation. Pain and joy, suffering 
and delight, adversity and triumph, and more particularly the bitter- 
sweet triumph of accepting adversity as part of our human lot: these 
are the ingredients, and we all mix them differently according to our 
manner of experience. 

Nevertheless there are some styles which seem able to balance them 
up better than others. What we call the baroque period of music had 
such a style, and the remarkable hold which this period has now 
gained on our affections is certainly not due to its having run away 
from one end of our human paradox by divorcing passion from serenity 
or keeping music unemotional. Not even Stravinsky (least of all 
Stravinsky) does this, although he seems to think he does. The hold 
which baroque music has gained over our affections is due precisely to 
the fact that its most typical composers, from the supreme case of 
J. S. Bach downwards, had each in their degree the secret of not only 
balancing the passion and serenity, but balancing them at a high level 
of intensity. Bach's music is very passionate and very serene, not in 
crude alternation, but as an integrated whole. If ever there was a case 
of transcending the opposites in a reconciliation which is more than 
the sum of its parts, that case is Bach's music. This is the measure of 
our problem in doing justice to him with our interpretation. 

Because the question of how emotional our interpretation of early 
music ought to be is a primary question which takes precedence even 
over the more complex question of how to apply the right kind of 
emotion in the right kind of way, I have collected a variety of evidence 
bearing on it. This is followed by a few further quotations to remind 
us of the tranquility which is the complementary aspect. I have then 



been able to find one or two In which the reconciliation of the paradox, 
normally so unconsciously achieved, has to some extent broken through 
into conscious description. I believe the reader may find a comparison 
of these passages helpful in establishing the underlying mood which is our 
best starting-point. (For fuller bibliographical detail see Bibliography.) 


(1) Charles Butler, Principles of Musik, London, 1636, p. 109, citing 

the fifth-century St. Augustine: 

*O how I wept at thy Hymns and Songs, being vehemently moved with 
die voices of thy sweet-sounding Church. Those Voices did pierce mine 
ears, and thy truth distilled into mine heart; and thereby was inflamed in 
me a love of Piety: the tears trickled down, and with them I was in 

happy case/ 

(2) William Prynne, Histriomastix, London, 1633, Ch. XX, citing the 
twelfth-century Bishop Ethelred: 

'Whence hath the Church so many Organs and Musicall Instruments? 
To what purpose, I pray you, is that terrible blowing of Belloes, expressing 
rather the crakes of Thunder, than die sweetnesse of a voyce? To what 
purpose serves that contraction and inflection of the voyce? This man sings 
a base, that a small meane, another a treble, a fourth divides and cuts 
asunder, as it were, certaine middle notes. One while the voyce is strained, 
anon it is remitted, now it is dashed, and then againe it is inlarged with a 
lowder sound. Sometimes, which is a shame to speake, it is enforced into 
a horse's neighings; sometimes, die masculine vigour being laid aside, it 
is sharpened into the shrilnesse of a woman's voyce; now and then it is 
writhed, and retorted with a certaine artificiall circumvolution. Sometimes 
thou may'st see a man with an open mouth, not to sing, but, as it were, 
to breathe out his last gaspe, by shutting in his breath, and by a certaine 
ridiculous interception of his voyce, as it were to threaten silence, and now 
again to imitate the agonies of a dying man, or the extasies of such as suffer/ 

(3) Baldassare Castiglione, The Courtyer, Venice, 1528, transl. Sir 
Thomas Hoby, London, 1561, Everyman ed. 1928, p. 61 : 

'Marke me musicke, wherein are harmonies sometime of a base sound 
and slow, and otherwhile verie quicke and of new devises, yet doe they all 
recreate a man, but for sundrie causes, as a man may perceive in the manner 
of singing that Bido useth, which is so artificall, cunning, vehement, 
stirred, and such sundrie melodies, that the spirites of the hearers move all 
and are inflamed, and so listing, a man would weene they were lift up into 

And no lesse doth our Marchetto Cara move in his singing, but 
with a more soft harmony, that by a delectable way and full of mourning 



sweetcnes maketh tender and percetli the mind, and sweetly imprinteth 
in it a passion full of great delite/ 

(4) Myles Coverdale, Goostly Psalmes, London, [? 1539], preface 
'Unto the Christen reader* : 

* Yf yonge men also that have the gyfte of syngynge, toke theyr pleasure 
in soch wholsome balettes ... it were a token, both that they felt some 
sparks of Gods love in theyr hertes, and that they also had some love unto 
hym, for truly as we love, so synge we: and where our affeccyon is, 
thence conuneth our myrth and joye . . .* 

(5) Charles Butler, Principles of Musik, London, 1636, p. 92: 

'[Good composing is impossible] unless the Author, at the time of 
Composing, be transported as it were with some Musical fury; so that 
himself scarce knoweth what he doth, nor can presently give a reason for 
his doing.' 

(6) Samuel Pepys, Diary, 27 Feb., 1667/8: 

'That which did please me beyond any thing in the whole world was 
the wind-musick when the angel comes down, which is so sweet that it 
ravished me, and indeed, in a word, did wrap up my soul so that it made 
me really sick, just as I have formerly been when in love with my wife; 
that neither then, nor all the evening going home, and at home, I was able 
to think of any thing, but remained all night transported.' 

(7) Angelo Berardi, Ragionamenti Musicali, Bologna, 1681, p. 87: 
'Music is the ruler of the passions of the soul.' 

(8) Frangois Raguenet, Comparison between the French and Italian 
Music, 1702 ? transl. ? J. E. Galliard, 1709, ed. O. Strunk, Mus. Quart., 
XXXII, 3, July, 1946, p. 422: 

'[Music is] transport, enchantment and extasy of pleasure . . .' 
[Translator's f.n. 9 p. 419:] 1 never met with any man that suffered his 
passions to hurry him away so much whilst he was playing on the violin 
as the famous Arcangelo Corelli, whose eyes will sometimes turn as red 
as fire; his countenance will be distorted, his eyeballs roll as in an agony, 
and he gives in so much to what he is doing that he doth not look like the 
same man/ 

(9) Dr. Charles Burney, Present State of Music in Germany, London, 
1773, II, p. 269: 

'[C.P.E. Bach at the clavichord] grew so animated and possessed, that he 
not only played, but looked like one inspired. His eyes were fixed, his 
underlip fell, and drops of effervescence distilled from his countenance. 



He said, If he were to be set to work frequently, in this manner, he should 

grow young again.' 

(10) C. H. Biainville, V Esprit de I 9 Art Musical, Geneva, 1754, p. 14: 

*A musician falls into inspiration at the moment when he least thinks of 
it; his imagination is fired, his heart swells, his blood pulses rapidly 
without his volition; a luminous cloud surrounds him, he is transported 
into a vast space; it is that which has reality, all his senses lend him their 
mutual aid, and are transformed piece by piece into passion, the image 
which he desires to paint; it all comes pressing upon him, he guides and 
selects. Exalted above himself, he traces, without knowing it, the beauties 
which he scarcely understands : like a second Pythian, he falls into a frenzy, 
he speaks the language of the Gods; he is drained at last, the forces fail him; 
he returns to himself . . / 

(11) [J. Mainwaring] Memoirs of. . . Handel, London, 1760, p. 52: 

'The audience was so enchanted with this performance, that a stranger 
who should have seen the manner in which they were affected, would 
have imagined they had all been distracted.' 


(12) Richard Allison, Howres Recreation in Musicke, London, 1606, 
foreword, citing the early sixteenth-century founder of the Reformation: 

'Musicke, saith he [Martin Luther] to Divels we know is hateful and 
intolerable, and I plainely thinke, neither am I ashamed to averr it, that 
next Theologie, there is no Arte comparable with Musicke: for it alone 
next to Theologie doth affect that, which otherwise only Theologie can 
perforate, that is a quiet and a cheareful minde/ 

(13) Thomas Sternhold, Psalter, London, 1560, title-page: 

'Very mete to be used of all sorts of people privately for their godly 
solace and comfort, laiying aparte all ungodly songues and ballades which 
tende only to the nourishing of vice, and corrupting of youth . . / 

(14) John Milton, Of Education, London, 1644, ed. of 1836, p. 162: 

'The solemn and divine harmonies of Music . . . have a great power 
over dispositions and manners, to smoothe and make them gentle from 
rustic harshness and distempered passions.' 

(15) Jean- Jacques Rousseau, Dictionnaire de Musique, Paris, 1768, 
s.v* 'Imitation' : 

'Night, sleep, solitude and silence are among the great pictures of music.' 




(16) Sir Thomas Browne, Rellgio MedicL London, 1542, Everyman 
ed., p. 79: 

"It is my temper, and I like it the better, to affect ail harmony, and sure 
there is musick even in the beauty, and the silent note which Cupid 
strikes, far sweeter than the sound of an instrument. For there is a musick 
where ever there is a harmony, order or proportion . . . 

'Whosoever is harmonically composed delights in harmony; which 
makes me much distrust the symmetry of those heads which declaim 
against all Church-Musick. For my self, not only from my obedience, but 
my particular Genius, I do embrace it: for even that vulgar and Tavern- 
Musick, which makes one man merry, another mad, strikes in me a deep 
fit of devotion, and a profound contemplation of the First Composer. 
There is something in it of Divinity more than the ear discovers: it is an 
Hieroglyphical and shadowed lesson of the whole World, and creatures 
of God; such a melody to the ear, as the whole World, well understood, 
would afford the understanding. In brief, it is a sensible fit of that harmony 
which intellectually sounds in the ears of God.' 

(17) Thomas Mace, Mustek's Monument, London, 1676, p. 19: 

'But when That Vast-Conchor ding-Unity of the whole Congregational- 
Chorus, came (as I may say) Thundering in, even so, as it made the very 
Ground shake under us; (Oh the unutterable ravishing Soul's delight!) In the 
which I was so transported, and wrapt up into High Contemplations, that 
there was no room left in my whole Man, viz. Body, Soul and Spirit, for 
any thing below Divine and Heavenly Raptures . . . 

[p. 118] *. . . Musick speaks so transcendently, and Communicates Its 
Notions so Intelligibly to the Internal, Intellectual, and Incomprehensible 
Faculties of the Soul; so far beyond &&. Language of Words, that I confess, 
and most solemnly affirm, I have been more Sensibly, Fervently, and 
Zealously Captivated, and drawn into Divine Raptures, and Contemplations, 
by Those Unexpressible, Rhetorical, Uncontroulable Persuasions, and Instruc- 
tions of Musicks Divine Language, than ever yet I have been, by the best 
Verbal Rhetorick, that came from any Mans Mouth, either in Pulpit, or 

'Those Influences, which come along with It, may aptly be compar'd, 
to Emanations, Communications, or Distillations, of some Sweet, and 
Heavenly Genius, or Spirit; Mystically, and Unapprehensibly (yet Effectually) 
Dispossessing the Soul, and Mind, of All Irregular Disturbing, and Unquiet 
Motions; and Stills, and Fills It, with Quietness, Joy, and Peace; Absolute 
Tranquillity, and Unexpressible Satisfaction.' 



The Interplay of Styles 


With sixteenth-century polyphony, the basic problem is to keep the 
seamless tapestry of the counterpoint intact while bringing out the 
themes and entries with sufficient light and shade. In a Monteverdi 
opera, the basic problem is to mould a single expressive monody to 
each fluctuation of feeling, words and drama, with the support of 
simple but extraordinarily intense and significant harmony. In the main 
baroque period, the basic problem is not quite either of these; it is 
something between the two. 

We have to build up a tremendously strong, balanced tension between 
a melody less fluid than Monteverdi's, and a bass less static. The 
middle is usually the least important, or at any rate the least well 
defined. Even the favourite trio-sonata form, with its twin melodies 
and its strongly melodic bass-line figured for the accompanist's impro- 
visations, has more the nature of an upper part going into duplicate 
than of fully spaced-out counterpoint. The fugal form itself is no longer 
equal-voiced polyphony in quite the old skteenth-century meaning; it, 
too, consists of progressions on a harmonic bass, however skilfully 
embodied in a contrapuntal texture. And in spite of the opposition of 
solo and ripieno in concertos, in spite of the fashionable echo-effects 
inherited from the brilliant Gabrieli tradition, the contrasts of baroque 
music are not characteristically within the movement. They are between 

That is the principle of the suite. Most baroque 'sonatas' are really 
suites, with movements which are either monothematic or thematically 
homogeneous. Even opera lost much of its dramatic flexibility as the 
direct influence of Monteverdi waned; a typical baroque opera has 
something of the quality of a protracted vocal suite. In music con- 
structed, in this characteristic baroque fashion, on the suite principle, 
we have to build up a further tension, by every variety of expressive 
resource, between the contrasting movements. 

We have here, then, two fairly general guides to baroque interpreta- 
tion: the strong balance between the top and bottom of the music, 
which implies keeping the bass-line nearly or quite as telling as the 
upper melody or melodies; and the strong opposition between move- 



ments, which Implies somewhat sustained moods not unnecessarily 
diversified within the movement. 

This is, of course, the merest starting-point. As such, these two 
broad rules are, in my experience, very well worth bearing in mind; 
but they are obviously subject (especially the second of them) to 
numerous and important exceptions; and there are many other factors 
to be taken into account, among which the date, the national style and 
the immediate purpose of the music are among the most conspicuously 


(18) J. S. Bach, memorandum to Leipzig Town Council, 1730 (in 
Spitta's Bach, transl. Bell and Fuller-Maitland, 1889): 

'The present status musices is quite different to what it used to be for- 
merly the art being much advanced and taste marvellously changed, so 
that the old-fashioned kind of music no longer sounds well in our ears.' 

(19) Joachim Quantz, Essay, Berlin, 1752, XVIII, 6: 

'The old musicians complain of the melodic extravagances of the 
young, and the young mock the dryness of the old/ 

(20) Dr. Charles Burney, Present State of Music in Germany, 
London, 1773, II, p. 156: 

* Quantz' music is simple and natural; his taste is that of forty years ago/ 

Thus J. S. Bach regarded himself as a modernist; and his music 
shows that he was in cordial sympathy with the then modern 'galant* 
tendencies in general, and with French idioms in particular. Quantz, 
who was also in sympathy with French music, was something of a 
conservative; Burney (op. cit. 9 p. 82), called his great Essay of 1752 
'classical'. Though this Essay was not actually published in J. S. Bach's 
lifetime, its whole tendency suggests that its contents are broadly 
applicable to J. S. Bach's music. (I have used freely the French edition.) 

J. S. Bach's most famous son, C. P. E. Bach, published the first part 
of his still more distinguished Essay in 1753, only a year later. He too, 
like his father, was a modernist, but there is a great difference. The 
foundation of J. S. Bach's art was traditional in a degree in which 
C. P. E. Bach's was not. Looking beneath the surface, and with two 
centuries of historical perspective to help us, we might say that both 
men were beyond their time, but the father in a different direction from 
that in which their time was then moving, and the son in the same 
direction. In C. P. E. Bach's own time, it merely appeared that the 
father was old-fashioned, and the son a leader of fashion. The son 



himself acknowledged his father's greatness and his own immeasurable 
debt to him; yet the tendency of C. P. E. Bach's Essay is somewhat 
further from 1 S. Bach than *Quantz' Essay, and its contents, though 
still largely applicable to J. S. Bach's music, are somewhat less generally 
so. Somewhat more caution is needed in arguing from C. P. E. Bach 
to J. S. Bach than in arguing from Quantz to J. S. Bach. 

This Is a characteristic situation. We cannot always count on finding 
evidence as close in time or place as we should wish, and slightly 
distant evidence may be better than a purely modern guess. The evidence 
leaves surprisingly few major uncertainties about the substance of the 
conventions for baroque music, but it is often extraordinarily hard to 
discover how widely they apply. We are not helped by the fact that no 
baroque authority displays a very sound or well-informed historical 
sense; most of them display none at all. On the other hand, their 
interest in national differences was unbounded, and here we have at 
least no dearth of evidence. 

The reader is recommended to Paul Henry Lang's definitive survey 
of the cross-currents of national influence in baroque music (Music in 
Western Civilization, New York, 1941, Ch. 10 onwards), but the 
following is a brief summary of the relationships most relevant to 
interpretation. 1 


(a) We may think of Italy as the radiating centre of baroque music. 
The foreign influences returning to that country were comparatively 
slight. Her exports included the new Italian monody of the turn of the 
sixteenth and seventeenth centuries, as well as the operatic forms which 
partly evolved from and partly succeeded to that highly novel species 
of musical declamation; these spread at different rates in different 
directions, but were paramount all over Europe by the eighteenth 

1 Before we go on, it may be as well just to recall that the term baroque, useful and even 
Indispensable though we now find it as a chronological label, was originally chosen for 
wrong reasons and along a wrong line of thought. 

The baroque period was not unacquainted with the word baroque. 'Hard, baroque and 
almost unsingable' (J.-J. Rousseau, 1753); 'baroque, that is, coarse and uncouth* (Burney, 
1773); A a baroque music is that which is filled with modulations too hard on the ear, and of 
which the movement is awkwardly interrupted' (Meude-Moupas Dictionary, 1787). In our 
day, the O.E.D. gives 'irregularly shaped; grotesque; odd; spec, of a florid style of late 
Renaissance architecture prevalent in the eighteenth century'; Larousse gives 'un style 
bizzarre.' What a word to have to use of J. S. Bach! 

The line of thought was that if this architecture is florid, so is much of the music. St, 
Paul's Cathedral can be caUed florid, Le. baroque, by those who do not like it. To those 
who do, it can still seem as essentially secular as Voltaire, whereas Bach is as spiritual (in 
some sense hard to define) as Winchester or Salisbury. Luckily hardly anyone now remem- 
bers the line of thought. We shall all go on using the label, but we add to our confusion and 
not to our understanding if we think of it as meaning anything. The only things it could 
mean are things untrue, uncouth, bizzarre and in a word, baroque. 



century. The same in a lesser degree is true of the bold contrasts of 
sonority and texture (some of the most massive effects In baroque 
music) particularly associated with the Gabrieli school. The great 
Italian organ-school of Frescobaldi, with his powerful structure yet 
brilliant fluency, had most effect in Germany, where it was joined by 
the equally valuable influence of the early seventeenth-century English 
organists and virginalists (John Bull, above all) by way of the Italianised 
Netherlander Sweelinck, and other channels. The Italian violinist- 
composers, like the Italian singers, gained a supremacy seriously chal- 
lenged only in France. The highest extent of the Italian leadership lasted 
a full three-quarters of a century, from about 1675 to about 1750, and 
it was not much lower in the decades before and after. In other fields, 
however, particularly keyboard music and orchestral suites, the French, 
too, were an internationally exporting nation over much of the same 
period, especially the latter part of it. 

(b) It is the paradox of the French national style that its most 
classical exponent, Lully, was himself an Italian-born immigrant. Even 
before Lully, the stimulus to French opera came as much from some 
brief visits by Italian companies as from the indigenous ballets de com. 
But Lully, in the mid-seventeenth century, acclimatised himself with 
uncanny skill, studying every subtlety of the French language to become 
one of the greatest masters in setting it to music. The rhythms and the 
cadences of spoken French entered not only into Lully's vocal tech- 
nique, but into his entire melody. Once acclimatised, Lully made no 
concessions to the Italian taste. It was French-born composers, among 
them Couperin the Great himself early in the eighteenth century, who 
added to their output of French-style suites a number of sonatas on the 
Italian model to which, therefore, Italian and not French conventions 
of interpretation should largely be applied. 

(c) Germany was hardly a musically exporting country at all during 
the baroque period, though the German violinist-composers, with their 
typical predilection for solo polyphony such as culminated in Bach's 
six astonishing unaccompanied suites, were a partial exception from an 
early date in the seventeenth century onwards. For the rest, the Italian 
trio-sonata was accepted, but with an interesting and markedly trans- 
forming disposition to replace two violins and a bass by a violin, a 
gamba and a bass, which not only deepens the sonority but distributes 
the part-writing more evenly in pitch. There were a number of English 
viol-players in Northern Germany in the seventeenth century, but so 
far from introducing the splendid English string fantasies of their day, 
they merely contributed to the rather uninteresting chamber dance 
suites then current in their adopted country. 

Late in the seventeenth century, French orchestral, chamber and 



solo suites were imported into Germany with growing enthusiasm, and 
at the same time many of the French conventions of interpretation, for 
which, therefore, we have to watch from this time onwards. J. S. Bach 
himself, though very open to the Italian influence (above all of Vivaldi) 
was also a profound admirer of the French idioms, however much he 
modified both influences in adapting them to his Germanic background 
and his own unique personality. His keyboard music other than for 
organ is mainly, though not entirely, founded on French models and 
in need of a French style of interpretation. His organ music is basically 
indigenous, with that solid power combined with rich poetic feeling 
which was the peculiar inheritance of Protestant German music a 
deeply religious inheritance in which the Lutheran chorale played a 
crucial role, not only as a symbol but also in conditioning the form. 

(d) England in the first half of the seventeenth century became 
musically almost as much of an island as she is geographically. She had 
already imported the Italian madrigal, and some elements of the 
Italian string fantasia, out of both of which she proceeded to manu- 
facture highly indigenous products. Their structure was by then ex- 
tremely old-fashioned on continental standards, being genuinely 
polyphonic in the sixteenth century meaning of the word; their harmony 
was advanced and grew more so, with a richer vein of dissonance than 
any contemporary Italian. None of this, however, was exported in turn, 
although the very expressive and brilliant keyboard music of the 
beginning of the century had, as we have seen, passed into the conti- 
nental tradition which led eventually to Bach and with it some 
advanced elements of fingering technique. 

By the middle of the seventeenth century, the Stuart viol fantasies, 
one of the great schools of early chamber music, had reached their 
height, and were nearing their end, with no trace of foreign influence 
since the initial cross-fertilisation from Italy (they had English roots as 
well). A few attempts to introduce Italian monody had to compete 
with another form of monody, indigenous, quite as expressive in its 
different way, and in a certain sense more maturely perfected, by 
Dowland and the other lutenist song-composers of solo ayres. There 
was some mutual interchange with France, mainly in lute music, airs 
de cour, and light dance forms: English lute ornamentation is thought 
to have inspired the French lutenists and through them the French 
keyboard school; but none of this was on a very substantial scale. 

At the Restoration, however, this relative isolation ended. The 
French influence was now cultivated in England both deliberately and 
extensively. Charles II, with his French upbringing during his exile, 
found the traditional English polyphony distasteful, and took active 
steps to replace it both in sacred and secular music with a more con- 



genial vein to which, it must be admitted, the English musicians took 
with alacrity. The result was English enough, since in the later seven- 
teenth century English musicians still had the genius, as they no longer 
had on Handel's arrival in the early eighteenth, to assimilate a foreign 
influence and make it their own. But the need to introduce some of the 
French conventions of interpretation is plain. This is very markedly 
the case with Purcell, whose sacred and operatic music alike is con- 
spicuous for its French idioms or rather for what would be French 
idioms if they were not so astonishingly transmuted by that most 
individual of English composers. Their performance needs French 
elements, at any rate, particularly with regard to rhythm. In his youthful 
string fantasies, Purcell is wholly English; in his trio-sonatas, he is 
deliberately using Italian models and half-unconsciously including 
French models, so that conventions from both of those national 
traditions are needed in the interpretation. Their mood and harmony 
are very English and very original: no such searching harmony as this 
was appearing on the continent in PurcelTs lifetime. His music had a 
certain limited but unmistakable influence on Handel. Apart from that, 
the English tradition then ran very thin and weak till Elgar's genius 
restored it to greatness. 

(e) This whole problem of national styles in the interpretation of 
baroque music is the subject of a valuable discussion by Thurston Dart 
in his short but brilliant Interpretation of Music (London, 1954, pp. 
77 ff.), where among other suggestions he gives a simple hint remarkably 
effectual in nine cases out of ten. He points out that we have usually 
a clue to the national style in question, if we notice what language is 
used for any title or expression mark given to the piece. Thus suite, 
ordre, allemande, courante, gigue, sarabande, vivement, vite, lentement, 
etc. normally imply a French idiom. Partita, sonata, allemanda, 
coranto, giga, sarabanda, allegro, presto, grave, adagio, etc,, normally 
imply an Italian idiom. 

In the case of the courante, the name is a less reliable guide than 
usual, partly because different spellings confuse the issue of language 
(this does occur with other titles also) and partly because the Italian 
form is found quite frequently with the French name. But here the form 
itself is readily distinguishable: the French in a characteristic alterna- 
tion of simple and compound triple rhythms (e.g., 3-2 and 6-4); the 
Italian in simple triple rhythm throughout or almost throughout, and 
a more smoothly flowing movement. Couperin sometimes pairs the 
two kinds by way of contrast. The saraband is another interesting case: 
there is a very light and rapid English seventeenth-century variety; a 
more lyrical Italian variety much favoured by Corelli; and a still more 
lingering and tender French variety, adopted and carried often to great 



lengths of embellishment by J. S. Bach. The dance itself was thought 
to come from Spain, and perhaps before that from the Near East; but 
such attributions are notoriously uncertain. It must be remembered, 
too, that as musical structures, dance-forms are apt to develop very far 
from their folk or ballroom origins. It is their nationality as music, not 
as dances, which affects their interpretation in the developed state. 


(21) M. Mersenne, Harmonie Universelle, Paris, 1636-7, II, vi, 356: 

4 As to the Italians, they . . . represent as much as they can the passions 
and the affects of the soul and the spirit; for example, anger, fury, spleen, 
rage, faintheartedness, and many other passions, with a violence so 
strange, that one judges them as if they were touched with the same 
affects as they represent in singing; in place of which our Frenchmen are 
content to caress the ear, and use nothing but a perpetual sweetness in 
their songs; which hinders their energy [a detailed study of the means 
employed by composers follows].' 

(22) Georg Muffat, Florilegium /, Augsburg, 1695, preface: 

'[The French style is now catching on in Germany and Muffat himself 
is a pupil of Lully ; this style has] natural melody, with an easy, and smooth 
tune, quite devoid of superfluous artifices, extravagant divisions [impro- 
visatory embellishments], and too frequent and harsh leaps/ 

(23) Frangois Raguenet, Comparison Between the French and Italian 
Music, Paris, 1702, transl. ? Galliard, London, 1709, ed. Strunk, 
Musical Quarterly, XXXII, 3, pp. 417^.: 

"The French, in their airs, aim at the soft, the easy, the flowing and 
coherent .... The Italians venture at everything that is harsh and out of 
the way, but then they do it like people that have a right to venture and 
are sure of success ... as the Italians are naturally much more brisk than 
the French, so are they more sensible of the passions and consequently 
express 'em more lively in all their productions. If a storm or rage is to be 
described in a symphony, their notes give us so natural an idea of it that 
our souls can hardly receive a stronger impression from the reality than 
they do from the description; everything is so brisk and piercing, so 
impetuous and affecting, that the imagination, the senses, the soul and the 
body itself are all betrayed into a general transport; 'tis impossible not to 
be borne down with the rapidity of these movements. A symphony of 
furies shakes the soul; it undermines and overthrows it in spite of all its 
care; the artist himself, whilst he is performing it, is seized with an unavoid- 
able agony; he tortures his violin; he racks his body; he is no longer 



master of himself, but is agitated like one possessed with an irresistible 

[p. 415] "our masters [in France] touch the violin much finer and with a 
greater nicety than they do in Italy.' 

(24) Le Cerf de la Vieville, in Histoire de la Musique of 1725, II, p. 61 : 

'Our violins [in France] are more tranquil than there in Italy . . . [the 
same work, IV, p. 149] : The Italian instrumentalists, as much for accom- 
paniment as in. their pieces, have no other merit than that ot drawing 
plenty of sound from their instruments.' 

(25) Roger North, Autobiography, c. 1695, ed. Jessopp, London, 
1887, sect. 114: 

'The court [of Charles II] about this time [later 17th century England] 
entertained only the theatrical music and French air in song, but that 
somewhat softened and variegated; so also was the instrumental more 
vague and with a mixture of caprice. . . , At length the time came off 
the French way and fell in with the Italian, and now that holds the ear. 
But still the English singularity will come in and have a share/ 

(26) Joachim Quantz, Essay, Berlin, 1752, X, 19: 

* After the pupil has gained also a general idea of the difference of taste 
in Music, he must recognise distinctive pieces of different nations and 
provinces, and learn to play them each according to its kind. . . . The 
diversity of distinctive pieces is commoner in French and German music, 
than in the Italian and in that of the other nations. Italian music is less 
restrained than any other; but the French is almost too much so, whence 
it comes about perhaps that in French Music the new always seems Hke 
the old. Nevertheless the French method of playing is not at all to be 
despised, above all an Apprentice should be recommended to mix the 
propriety and the clarity of the French with the chiaroscuro of the Italian 
instrumentalists. It is mainly a question of bow-strokes and the use of 
ornaments, of which last the Italian instrumentalists make too many, and 
the French in general too few [he is presumably speaking here of impro- 
vised free embellishments, not of specific ornaments, of which the French 
often wrote, and performed, a maximum]. 

c [XVIII, 53: There are two leading musical nations.] These are the 
Italians and the French. Other nations are ruled in their taste by these two. 
, , . [76: Italian instrumentalists are decadent and eccentric, but skilful. 
Italian singing is supreme]. In a word: Italian music is arbitrary and 
French is limited: so that if the effect is to be good, the French depends 
more on the composition than on the performance, the Italian almost as 
much, and in some pieces almost more, on the performance than on the 



'[XVII, vii, 14: it is disastrous] to make up an orchestra of people some 

of whom play in the Italian style, others in the French, and others again in 
another style/ 

(27) J. Cazotte, Observations sur la lettre del.- J. Rousseau [? Paris], 
1753, p. 11: 

*[The French airs de Ballet are] full of fire, of variety, of gaiety, and of 


(28) E.G. Freron, Lettre sur la Musique Frangoise [Paris, 1753] : 

'The extreme vitality [of the Italians inclines them] even to an excess of 
imagination. . . . But we [French] who are used to serious, tender and 
sustained passions ... are moved by the same things in a different way 
from the Italians/ 

(29) C. C. de Rulhiere, Jugement de F Orchestra de F Optra [Paris, 

Italian music is full of epigrams; French music is more noble.' 

(30) [L'Abbe Laugier] Apologie de la Musique Frangoise [Paris] 1754, 
p. 5: 

'The character of a national music does not depend at all on the quality 
of the language; but on the amount of genius. It is genius, and genius 
alone which gives birth to whatever music has that is most pleasant and 
moving. Its tender sweetness, its light vivacities, its sad and sombre langors, 
its harshnesses, its furies, its swiftnesses, its confusions, are the fruit, not of 
a language more or less favourable to the charms of melody; but of a 
spirit which flows freely into inventions full of fire, and which subordi- 
nates the harmony to its ideas. . . . Every nation where genius can light 
its flame, can have true music.' 

To recognise the main varieties of national style, even when found 
elsewhere than in their countries of origin, is not, perhaps, so difficult 
as it sounds. Once, for example, the salient characteristics of the French 
overture style are grasped at all, with its notated dots all meant to be 
double-dotted and its typical up-beat scales all delayed, quickened and 
taken at the rush, these characteristics are very recognisable whether 
they are found in the German-born Handel, the English Purcell, or 
even on occasion the Italian Vivaldi; and it soon becomes second 
nature to interpret them in the pointed French style (for which see 
Ch. XLIV below). Similarly with the four-square Italian opening 
allegro style; once it is known, there is no risk of confusing it with the 
jerkier French movement. There are, indeed, other national character- 
istics much less clearly distinguished than these but then they are also 
less important to distinguish. 




(31) John Cosyn, Psalmes, London, 1585, Dedication: 

\ . .for the private use and comfort of the godlie* in place of many 
other Songs neither tending to the praise of God, nor concerning any 
thing fit for Christian eares/ 

(32) The brothers Lawes, Choice Psalmes, London, 1648, preface: 

'. . . having been often heard, and well approv'd of, chiefly by such 
as desire to joyne Musick with Devotion . . / 

(33) John Playford, Introduction, London, 1654, eds, of 1672 on, 
preface 'Of Musick in General* : 

'The first and chief Use of Musick is for the Service and praise of God, 
whose gift it is. The second Use is for the Solace of Men, which as it is 
agreeable unto nature, so is it allowed by God as a Temporal blessing to 
recreate and Chear men after long study and weary labour in their 
vocations . . / 

(34) Thomas Mace, Mustek's Monument, London, 1676, p. 233: 

'Know, That in my Younger Time, we had Musick most Excellently 
Choice, and most Eminently Rare; both for Its Excellency in Composition, 
Rare Fancy, and Sprightly Ay re; as also for Its Proper, and Fit Performances 

'We had for our Grave Musick, Fancies of 3, 4, 5, and 6 Parts [for viols] 
to the Organ; Interpos'd (now and then) with some Pavins, Allmaines, 
Solemn, and Sweet Delightful Ayres; all which were (as it were) so many 
Pathettical Stones, Rhetorical, and Sublime Discourses; Subtil, and Accute 
Argumentations; so Suitable, and Agreeing to the Inward, Secret, and Intellectual 
Faculties of the Soul and Mind; that to set Them forth according to their 
True Praise, there are no Words Sufficient in Language; yet what I can best 
speak of Them, shall be only to say, That They have been to my self, (and 
many others] as Divine Raptures, Powerfully Captivating all our unruly 
Faculties, and Affections, (for the Time] and disposing us to Solidity, Gravity, 
and a Good Temper, making us capable of Heavenly, and Divine Influences . . . 

'And These Things were Performed, upon so many Eqtial, and Truly- 
Scizd Viols', and so Exactly Strung, Tuna, and Playd upon, as no one Part 
was any Impediment to the Other; but still (as the Composition required) by 
Intervals, each Part Amplified, and Heightned the Other; The Organ Evenly, 
Softly, and Sweetly Acchording to All . . . 

'But when we would be most Ayrey, Jocond, Lively, and Spruce; Then 
we had Choice, and Singular Consorts, either for 2, 3, or 4 Parts, but not to 
the Organ (as many (now a days) Improperly, and Unadvisedly perform such 
like Consorts with) but to the Harpsicon . . .' 



(35) P. F. TosL Observations on the Florid Song, Bologna, 1723, 
transl. Galiiard, London, 1742, p. 92: 

'By the Ancients . . . Airs were sung in three different Manners; for die 
Theatre, the Stile was lively and various; for the Chamber, delicate and 
finished; and for the Church, moving and grave. This Difference, to very 

many Moderns, is quite unknown. 

(36) Joachim Quantz, Essay, Berlin, 1752, XVII, vii, 53: 

4 With Church music . . . the expression as well as the tempo should be 
more moderate than in opera . . . 

[XVII, vii, 12] 'Church music requires a grander and more serious air 
than theatre music, and this last allows more liberty.' 

(37) [De Rochement] Reflexions sur r Opera Francois et . . . Italien, 
Lausanne, 1754, p. 90: 

'The language of music is very vague, and very indeterminate, we have 
already said: the most powerful, vehement, excessive image, the most 
highly charged, if you like, with passion, will be better, more just, more 
perfect, more pleasing, more free from musical defect, than the image of a 
moderate passion which seems too restricted and always a little cool. 
That is the true reason for the decided superiority of the Italians in 
instrumental music of a grand character. The same principle applied to 
vocal music gives a different result. It is unsuitable then to charge the 
image too much with passion, because it is necessary to have regard to the 
relative, and to confine oneself to them.' 

(38) J. P. Kirnberger, Kunst des reinen Satzes, Berlin, Part I, 2nd 
ed., 1774, Sect. V, p. 80: 

'The former, or strict, style is chiefly used in church music, which is 
always of a serious or solemn character; but the latter is chiefly proper to 
the stage and to concerted music, of which the purpose is to please the ear 
rather than to arouse serious or solemn feelings. It is therefore generally 
known as the galant style, and a number of elegant licenses and various 
departures from the rules are allowed to it/ 


Like all antipathies, the musical antipathy of Italy and France had 
its undercurrent of sympathy, from which presently a new synthesis 
arose. This took the form of what became known, loosely enough, as 
the 'galant style'. The word 'galant' here really means 'polite', in the 
sense of 'polite conversation', and for much of the music this descrip- 
tion could not be bettered. In polite conversation, the deeper emotions, 
when they are mentioned at all, are brought in with a light touch. The 
best of such music glows with a genuine radiance; but it is the radiance 



of sensibility and not of passion (Fr. sensibilite; Ger. 
recurrent terms in this connection). Sensibility is feeling rendered 
elegant. Yet it is still feeling, and potentially capable of burning through 
its polite conventions. 

The easy-going Telemann, who made both Luliy and the Italian 
operatic melody his models, may stand very well for the shallower 
forms of galant music. Domenico Scarlatti, born though he was like 
J. S. Bach and Handel in 1685, was an altogether galant composer 
whose elegance adorns but does not conceal Ms depth of feeling. J, S. 
Bach himself shows far more of the galant influence than is generally 
realised, and the quotation from Heinichen at (40) below could easily 
apply to him. This and the following quotation from Quantz at (41) 
below almost make us feel as if the long hidden ideals of Monteverdi 
were re-emerging, to lead on to Gluck, Mozart and that classical 
Viennese tradition to which we are still the heirs. For all its facile 
aspects, it was galant music which achieved the crucial transition 
between baroque music and classical music. At the end of that road 
stood Beethoven. 

(39) Gabriel Guillemain, sonatas, Paris, 1743 : 

Six sonatas of four parts or Galant and Witty conversations between a 
Transverse Flute, a Violin, a Bass Viol and the Thorough Bass.' 

(40) J. D. Heinichen, General-Bass, Dresden, 1728, Introd., p. 24:" 

*And, oh, how wonderfully it pleases the ear, when, in an exquisite 
piece of church music or other composition, we hear how a master 
musician has tried from time to time, by his galant idioms and the close- 
ness of his expression to the verbal text, to move the feelings of his 
hearers and thus to achieve successfully the true aim of music. 

[Note:] 'Whence practical modern musicians are righdy inclined to 
depart from the unseasoned character of a too antique church style/ 

(41) Joachim Quantz, Essay, Berlin, 1752, XI, 16: 

'The performer should, so to speak, change to a different emotion in 
every bar.' 


This last quotation from Quantz at (41) above would suggest a 
certain risk of chaos if we did not know that the galant composers were 
also insistent on clarity of form, and were, in fact, engaged in pioneering 
what became the most strong yet balanced form of all those which the 
nineteenth century inherited : sonata form. 

A suite-form movement is essentially a monad; but a sonata-form 
movement is a duad. Its themes, though not necessarily two in number, 



are organised in two fields of force. They are first expounded in op- 
posing keys; next developed, often in relation to each other, through a 
whole drama of modulations; then recapitulated with their keys and 
their opposition reconciled. It is a perfect microcosm of common 
human experience. Tension, relationship, resolution are elements of life 
itself, and as such, of course, no innovations of the post-baroque 
period. It is the way in which sonata form deploys these elements which 
was in some genuine degree a novelty. Essentially this novelty consists 
in a greater acceptance of conflict within the movement, and not 
merely, suite-fashion, between the movements. In terms made famous 
in another context by Hans Keller, to the 'unity of contrasting move- 
ments' there was added a greatly increased 'unity of contrasting themes'. 

In sonata form, and indeed in post-baroque music generally, the 
expression is conditioned to a greater extent by dramatic contrasts 
within the movement; in suite form, and baroque music generally, it is 
conditioned to a greater extent by dramatic contrasts between the 
movements. This is one of the practical differences which the interpreter 
does well to bear in mind. There are, of course, dynamic and other 
contrasts which can properly be made, and should be made thoroughly, 
within the normal baroque movement; any unnatural restraint in this 
respect can only be harmful But it is a mistake to work up more 
internal contrasts of expression than the music itself genuinely suggests. 
Where the music itself remains emotionally on one level, we should be 
content to remain on one level of expression without seeking an arti- 
ficial variety. This is a situation which can occur in music of any period, 
but which is much more frequent in baroque than in post-baroque 

It should be noticed, however, that the operative principle is the 
same in either case. Whether the main contrasts in the structure of the 
music fall within or between the movements, it is still the structure 
which conditions the expression. In working out the expression for a 
baroque composition, where the composer will have left a far greater 
share in the responsibility for it to the performer than is nowadays the 
case, it is particularly important to refrain from building up big effects 
for their own sake. It is particularly important to listen quietly for the 
implications of the music to make its own requirements plain. But this 
is, of course, only a more extreme case of what any good interpreter 
will do even in music where the main lines of the expression have been 
laid down beforehand, as they are not laid down in baroque music, 
by the composer himself through a systematic use of expression marks. 
Matching the expression to the music is an axiom of interpretation 
under whatever circumstances. 



Music as Expression 


The belief, common to the galant composers, that music expresses not 
only emotion but specific emotions, was elaborated into a regular 
'theory of affects' at the time; but it was not a new belief (indeed it is 
at least as old as Plato). 

(42) Sir Thomas More, Utopia, Louvain, 1516, etc., Bk. II, Everyman 
ed.,p. 109: 

'For all their musike bothe that they [the Utopians] playe upon instru- 
mentes, and that they singe with mannes voyce dothe so resemble and 
expresse naturall affections, the sound and tune is so applied and made 
agreable to the thinge, that whether it bee a prayer, or els a dytty of 
gladiies, of patience, of trouble, of moumynge, or of anger: the fassion of 
the melodye dothe so represente the meaning of the thing, that it doth 
wonderfullye move, stirre, pearce, and enflame the hearers myndes.' 

(43) R. Hooker, Lawes, London, V, 1597, Everyman ed., p. 146: 

6 An admirable faculty which music hath to express and represent to the 
mind, more inwardly than any other sensible mean, the very standing, 
rising, and falling, the very steps and inflections every way, the turns and 
varieties of all passions whereunto the mind is subject; yea so to imitate 
them, that whether it resemble unto us the same state wherein our minds 
already are, or a clean contrary, we are not more contentedly by the one 
confirmed, than changed and led away by the other. In harmony the very 
image and character even of virtue and vice is perceived, the mind de- 
lighted with their resemblances, and brought by having them often iterated 
into a love of the things themselves. For which cause there is nothing 
more contagious and pestilent than some kinds of harmony; than some 
nothing more strong and potent unto good. And that there is such a differ- 
ence of one kind from another we need no proof but our own experience, 
inasmuch as we are at the hearing of some more inclined unto sorrow and 
heaviness, of some, more mollified and softened in mind; one kind apter 
to stay and setde us, another to move and stir our affections; there is that 
draweth to a marvellous and sober mediocrity, there is also that carrieth 
as it were into ecstasies, filling the mind with an heavenly joy and for the 
time in a manner severing it from the body.' 



(44) Thomas Morley, Plaine and Easie Introduction, London, 1597, 

p. 177 (an adaptation of Zarlino): 

It followeth to shew you how to dispose your musicke according to 
the nature of the words which you are therein to expresse, as whatsoever 
matter it be which you have in hand, such a kind of musicke must you 
frame to it. You must therefore if you have a grave matter, applie a grave 
kinde of musicke to it, if a merrie subject you must make your musicke 
also merrie. For it will be a great absurdirie to use a sad harmonic to a 
merrie matter, or a merrie harmonie to a sad lamentable or tragical 
dittie. You must then when you would expresse any word signifying 
hardnesse, crueltie, bitternesse, and other such like, make the harmonie 
like unto it, that is somewhat harsh and hard but yet so it offend not. 
Likewise, when any of your words shall expresse complaint, dolor, 
repentance, sighs, teares, and such like, let your harmonie be sad and 
doleful, so that if you would have your musicke signifie hardness, cruelty 
or other such affects, you must cause the partes proceede in their motions 
without the halfe note . . . but when you woulde expresse a lamentable 
passion, then must you use motions proceeding by halfe notes.' 

(45) William Byrd, Gradualia, I, London, 1605, transl, Fellowes, 
Byrd, London, 1936, p. 85: 

*. . . There is a certain hidden power, as I learnt by experience, in the 
thoughts underlying the words themselves; so that, as one meditates 
upon the sacred words and constantly and seriously considers them, the 
right notes, in some inexplicable manner, suggest themselves quite 

(46) Charles Butler, Principles ofMusik, London, 1636, p. 1: 

'Music . . . having a great power over the affections of the mind, by 
its various Modes produceth in the hearers various affects.' 

(47) Thomas Mace, Mustek's Monument, London, 1676, p. 118: 

'And as in Language, various Humours, Conceits, and Passions, (of All 
sorts) may be Exprest; so likewise in Musick, may any Humour, Conceit, 
or Passion (never so various) be Exprest; and so significantly, as any 
Rhetorical Words, or Expressions are able to do.' 

(48) Johann Mattheson, Neu-Eroffhete Orchestre, Hamburg, 1713, 
p. 161 and p. 167, transl. Cannon, Johann Mattheson, New Haven and 
London, 1947, p. 129: 

'[Opera is the best medium of all for expressing] each and every Affectus 
[since] there the composer has the grand opportunity to give free rein to 
his invention. With many surprises and with as much grace he there can, 
most naturally and diversely, portray love, jealousy, hatred, gentleness, 



impatience, lust, indifference, fear, vengeance, fortitude, timidity, magna- 
nimity, horror, dignity, baseness, splendour, indigence, pride, humility, 
joy, laughter, weeping, mirth, pain, happiness, despair, storm, tran- 
quillity, even heaven and earth, sea and hell, together with ail the actions 
in which men participate . . . 

'Through the skill of composer and singer each and every Affectus can 
be expressed beautifully and naturally better than in an Oratorio, better 
than in painting or sculpture, for not only are Operas expressed in words, 
but they are helped along by appropriate actions and above all interpreted 
by heart-moving music/ 

(49) Joachim Quantz, Essay, Berlin, 1752, XVIII, 24: 

'To judge operatic music reliably, it must be examined if the symphony 
or the overture has a true correspondence with the subject of the entire 
piece or the first act, or at least the first scene, and if it is capable of putting 
the hearers into the passion which prevails in the first act, whatever it be, 
tender or sad, gay, heroic, furious, etc., . . . if the composer has expressed 
the passions well in the way the subject requires; if he has distinguished 
them from one another well, and set each in its suitable context ... if 
he has kept up the main character of the piece from start to finish . . . 
finally if most of the audience are moved by the music and put into the 
passions represented . . / 

(50) F. W. Marpurg, Der critische Musicus an der Spree, Berlin, 2 
Sept. 1749: 

'All musical expression has an affect or emotion for its foundation. A 
philosopher when expounding or demonstrating will try to enlighten our 
understanding, to bring it lucidity and order. The orator, the poet, the 
musician attempt rather to inflame than to enlighten. The philosopher 
deals in combustible matter capable of glowing or yielding a temperate 
and moderate warmth. But in music there is only the distilled essence of 
this matter, the most refined part of it, which throws out thousands of the 
most beautiful flames, always with rapidity, sometimes with violence. 
The musician has therefore a thousand parts to play, a thousand characters 
to assume at the composer's bidding. To what extraordinary under- 
takings our passions carry us! He who has the good fortune at all to 
experience the inspiration which lends greatness to poets, orators, artists, 
will be aware how vehemently and diversely our soul responds when it is 
given over to the emotions. Thus to interpret rightly every composition 
which is put in front of him a musician needs the utmost sensibility and 
the most felicitous powers of intuition/ 

(51) C. P. E. Bach, Essay, Berlin, 1753, III, 13: 

*A musician cannot move others without himself being moved. He 
will have to feel all the emotions he hopes to call up in his audience, since 



by showing his own mood he will rouse a similar mood in the listener. 
Where the passage is languishing and sad, the performer must languish 
and grow sad. When it is lively and joyful, the performer must likewise 
put himself into the mood in question. He must especially perform this 
duty in music of which the nature is highly expressive, whether it is by 
him or another composer. In the latter event he must be sure to take on 
the feeling which the composer intended in writing it. ... 

'It will be realised from the many emotions which music depicts that 
the expert musician has need of special talents and the ability to use them 
judiciously. He must sum up his audience carefully, their response to the 
feelings expressed in his programme, the actual place and other added 
considerations. Nature has wisely endowed music with all manner of 
attractions so that everyone may share in enjoying it. Hence it falls to 
the performer to please to the best of his power every last manner of 

(52) Joachim Quantz, Essay, Berlin, 1752, XVIII, 28: 

'And since instrumental music has, without words, to express different 
passions and to carry the hearer from one to another, as well as vocal 
music; we can readily see that to do that, and supply it in the absence of 
word or human voice, the composer and he who performs the music 
must alike have a feeling soul, and one capable of being moved/ 

(53) C. H. Blainville, U Esprit de YAn Musical, Geneva, 1754, p. 86: 

'There are also purely instrumental pieces pursued with such truth that 
they seem to suggest words, notions of passion, imagery or painting; such 
seems to me the language of Tartini [Locatelli, Geminiani, Leclair, 
Corelli and Senaille are subsequently mentioned in the same connection], 
a true language of sounds . . / 

(54) John Trydell, Two Essays, Dublin, 1766, p. 97: 

'Musical Sounds do of themselves, that is to say, merely by the Sound, 
suggest our Ideas. Chiefly those of the Passions, as Desire, Aversion, Joy, 
Sorrow and the like/ 

(55) Jean- Jacques Rousseau, Dictionnaire de Musique, Paris, 1768, 
s.v. 'Imitation*: 

*The art of music consists in substituting for the insensible image of the 
object that of the emotions which its presence excites in the heart of the 
contemplator. Not only does it agitate the sea, animate the flame of con- 
flagration, make the streams to flow, the rain to fall and the torrents to 
swell; but it paints the horror of a terrible desert, darkens the walls of a 
subterranean prison, renders the air tranquil and serene, and sheds from 
the orchestra a new freshness through the groves. It will not represent 



these things directly, but it will excite in the soul the same emotions as 
one experiences in seeing them. 

'[s.v. 'Genie'] It paints pictures by sounds; it makes even silence speak; 
it renders ideas by feelings, feelings by accents; and the passions which it 
expresses, it stirs from the bottom of the heart.* 

(56) Joachim Quantz, Essay, Berlin, 1752, Introd., 17: 

'[But] if anyone wants to know, what ought to be the object of our 
researches and our reflections, here is our answer. After a new Composer 
shall have solidly learnt the rules of Harmony, which are only a matter of 
what is least and most convenient in Composition, although again many 
people are found without this knowledge; it is necessary that according 
to the purpose of each piece, he tries to make a just choice and a happy 
blend of his thoughts from start to finish; that he expresses properly die 
movements of the soul; that he sustains a flowing melody; that he applies 
himself to being original, and yet always natural in his modulation and 
just in his measure; that he sets above all light and shade; that he knows 
how to confine his inventions within reasonable bounds; that he does not 
make his melody stand still too much, and that he does not repeat the 
same thoughts too often; that he only composes in a manner suitable 
both to the voice and to the instruments; that in vocal Music he does not 
write against the scansion of the syllables, and still less against the sense of 
the words ; and finally that he tries to acquire a sufficient knowledge, both 
of the manner of singing, and of the different qualities of each instru- 
ment . . . 

'He who would be a Composer must have a spirit lively and full of 
fire, a soul open to tender feelings; a happy blend of what Philosophers 
call Temperament, in which not too much melancholy is found; much 
imagination, invention, judgment and discernment; a good memory; a 
sensitive and keen ear; an outlook good and shrewd and full of hu- 
mility . . / 

(57) [De Rochement] Reflexions sur F Opera Francois et sur r Opera 
Italien, Lausanne, 1754, p. 27: 

'Music only makes such a pleasant impression because it awakens the 
image of the passions/ 

(58) C H. BlainviUe, L 9 Esprit de FArt Musical, Geneva, 1754, p. 39: 
'A certain disorder which pleases . . .' 

(59) J. F. Marmontel, Essai sur les Revolutions de la Musique en 
France [Paris, 1777], p. 22: 

'The object of the arts which move the soul is not only emotion, but 
the pleasure which accompanies it. It is not enough for the emotion to be 
strong, it must also be agreeable.' 



(60) W. A. Mozart, Letter to his father, 26 Sept., 1781 (transl. E. 
Anderson, London, 1938): 

'For just as a man in such a towering rage [as Osmin in Die Entfuhrung 
aus dem Serail] oversteps all the bounds of order, moderation and propriety, 
and completely forgets himself, so must the music too forget itself. But as 
passions, whether violent or not, must never be expressed in such a way 
as to excite disgust, so music, even in the most terrible situations, must 
never offend the ear, but must please the hearer, or in other words must 
never cease to be music.' 


An illusion which can be very misleading to the modem performer, 
because it makes him think that he can improve on the original methods 
and puts him generally into a closed, patronising and sentimentalising 
state of mind, is the illusion that modern standards of technique are 
higher than early standards. That appears to be largely true in some 
departments, but the reverse of true in others. Our singers are sadly 
taxed by the virtuoso baroque parts. We have considerable trouble 
with early trumpet and horn parts, or again with J. S. Bach's unaccom- 
panied violin music (and he was not the man, nor was that the age, to 
have composed so much of It without adequate prospects of perform- 
ance). We have very few masters of the difficult baroque techniques of 
improvising continuo accompaniments or ornamental figuration. No 
doubt we read severe criticisms of contemporary performances in the 
early authorities; but so we do in the Times or the Telegraph today, 
the inference being not that our standards of performance are low but 
that our standards of criticism are high. Here, at any rate, are two 
baroque criticisms which cancel out (my italics). 

(61) Joachim Quantz, Essay 9 Berlin, 1752, Introd., 8: 

'[In the last century] how many changes have there not been in music in 
Germany? At how many courts, and in how many towns did it not 
flourish in the past, in a manner which brought up a good number of 
skilful people, and at present nothing but ignorance reigns in this respect. 
At the majority of courts which were once provided with skilful people, 
of whom a proportion were actually excellent, the custom is introduced 
at present, of giving the first places in the music, to people who would 
not deserve the last in a good orchestra . . .' 

(62) Dr. Burney, Present State of Music in Germany, London, 1773, 
ed. of 1775, II, p. 202: 

1 must observe that the musicians of many parts of Europe have dis- 
covered and adopted certain refinements, in the manner of executing even 
old music, which are not yet received in the Berlin school, where pianos 



and fortes are but little attended to, and where each performer seems try- 
ing to surpass his neighbour, in nothing so much as loudness . . . when a 
piece is executed with such unremitting fury, as I have sometimes heard, 
it ceases to be music.* 


(63) Sir Thomas Elyot, The Governour, London, 1531, Everyman ed., 
p. 25: 

\ . . some pleasaunt lernynge and exercise, as playenge on instruments of 
musicke, whiche moderately used and without diminution of honour 
... is not to be contemned.' 

(64) John Marston, The Malcontent, London, 1604, 1, iii: 

'Malevole Fool, most happily encountered: canst sing, fool? 
Passarello Yes, I can sing, fool, if you'll bear the burden; and I can 
play upon instruments, scurvily, as gentlemen do/ 

(65) Gervase Markham, Countrey Contentments, London, 1615, I, 

'[A gentleman] shou'd not be unskilful in Musick, that whensoever 
either melancholy, heaviness of his thoughts, or the perturbations of his 
own fancies, stirreth up sadness in him, he may remove the same with 
some godly Hymn or Anthem, of which David gives him ample examples.' 

(66) William Byrd, Psalmes, Songs and Sonnets, London, 1611, 

'Only this I desire; that you will be but as carefull to heare them well 
expressed, as I have beene both in the Composing and correcting of them. 
Otherwise the best Song that ever was made will seeme harsh and 
unpleasant, for that the well expressing of them, either by Voyces, or 
Instruments, is the life of our labours, which is seldom or never well per- 
formed at the first singing or playing. Besides a song that is well and 
artifically made cannot be well perceived nor understood at the first 
hearing, but the oftener you shall heare it, the better cause of liking you 
will discover: and commonly that Song is best esteemed with which our 
eares are most acquainted.' 

(67) Orlando Gibbons, Madrigals and Mottets, London, 1612, 

'Experience tels us that Songs of this Nature are usually esteemed as 
they are well or ill performed/ 

(68) Roger North, Autobiography, c. 1695, ed. Jessopp, London, 
1887, sect. 103: 

'Whereas in the country, where there is not such variety of gay things 
to call young folks away, it is incredible what application and industry 



will grow up in some active spirits, and this voluntarily, without being 

incited, as where there is an inlet to music, having time and a fancy to it, 
some will be wonderfully resigned to master a reasonable performance in 
it . . . I would not have families discouraged for want of a perfection, 
which, to say truth, is not to be had out of a trade/ 

(69) Joachim Quantz, Essay, Berlin, 1752, Introd., 2: 

* [Rules must first be given for would-be musicians] by which they can 
discern, whether they are provided with the qualities necessary to a 
perfect musician ... I speak here only of those who wish to make a 
profession of Music, and who count on excelling in it with time. Not so 
much is asked of those who only wish to take their pleasure in it [provided 
that being amateurs they do not give themselves the airs of professionals]/ 


(70) Giulio Caccini, Nuove Musiche, Florence, 1602, preface, transl. 
Playford, Introduction, London, 1654; ed. of 1674, p. 41 : 

'Art admitteth no Mediocrity, and how much the more curiosities are 
in it, by reason of the excellence thereof, with so much the more labour 
and love ought we, the Professors thereof, to find them out/ 

(71) Robert Dowland, Varietie of Lute-lessons, London, 1610, 

"Perfection in any skill [in this case lute-playing] cannot be attained 
unto without the waste of many yeares, much cost, and excessive labour 
and Industrie/ 

(72) Roger North, MS notes in Brit. Mus., early 18th cent., Add. 
325S6,/. 14: 

'Musick demands not only utmost spirit, and decorum in the com- 
position, but little less than perfection in the performance, which is not 
alwaies found/ 

(73) Joachim Quantz, Essay, Berlin, 1752, preface: 

*A man who only aspires to become a good Kipienist has not need of 
such great talents .... But he whose soul is not open to any feeling, 
who has clumsy fiagers, and no ear for music, will do better to apply 
himself to quite another science than Music . . / 

*To excel! in music, it is necessary to have an insatiable ambition, love 
and desire, to spare no pains and labour, and to have the courage to 
support all the discomforts which meet one in this kind of life ... a 
noble obstinacy . . . which inspires us to perfect ourselves more and 




It is only because we human beings are alike at bottom that com- 
poser, interpreter and listener (each with their own personal back- 
ground, great or little familiarity with the art, etc.) can come together 
in the shared experience of music. Yet on a less archetypal plane we 
are full of divergencies. It is these divergencies which give our individual 
interpretations their intrinsic value, and make it worth listening not 
just to one good interpretation but to many. 

(74) Joachim Quantz, Essay, Berlin, 1752, preface: 

1 seem at times to speak rather like a Dictator, resting what I suggest 
on a simple: it must, without advancing other proofs. 

'But it will be taken into account that it would be too long and at 
times impossible, to give demonstrative proofs on matters which nearly 
always look only to taste. Whoever does not want to trust mine, which I 
have endeavoured to purify by a long experience and by plenty of reflec- 
tions that I have made thereon, will be at liberty to try the contrary, and 
to choose then that which appears to him the best . , .' 

This confirms that it is unrealistic to ask: 'how exactly did J. S. Bach 
mean such-and-such a passage to be interpreted ?' There are no exact 
interpretations, and there are no unchanging ones not even from the 
same performer. There are only individual interpretations within the 
flexible though not indefinitely elastic boundaries of style. It is more 
realistic to ask: 'how approximately might the passage have been 
interpreted on some typical occasion by some good contemporary 
performer?' Since good performers varied even more widely in their 
interpretations in the baroque age than they do today, this should 
leave margin enough for any reasonable modern taste, whether leaning 
towards the driest and most correct or towards the warmest and most 
romantic rendering which the style will tolerate. 

(75) Roger North, MS. notes in the Brit. Mus., early 18th cent., 
Add. 32536, /. 14: 

'The taste of the audience is commonly prejudicate and Bizearre, some 
affect one kind, and will not hear another, and few allow any to be good 
that jumps not with their caprice.' 

(76) Joachim Quantz, Essay, Berlin, 1752, XVIII, 7-8: 

'Most people let themselves be ruled by ignorance, prejudice and 
passion. . . . Some like what is majestic and lively, others what is sad and 
sombre, and others what is tender and gay. The diversity of taste depends 
on the diversity of temperaments . . . one is not always in the same mood.* 







Accidentals in Early Music 


We have a problem over the accidentals in early music inherited from 
the more fundamental and historic problem of adapting the originally 
melodic modes first to harmony (a medieval achievement) and then to 
modulation (the great late-renaissance contribution). 

Western music since antiquity has virtually confined itself to the 
twelve semitones of the chromatic scale (crudely, the notes of the 

The modes further confined themselves to the five tones and two 
semitones of the diatonic scale (crudely, the white notes of the piano). 
Each mode distributes its two semitones in a different relationship to 
its centre of tonality ('final'): an absolute difference irrespective of 
context. Any chromatic modification of this distribution weakens the 

The keys take their choice of diatonic scale. But each key distributes 
its two semitones in the same relationship to its centre of tonality 
('tonic') : its difference is established by context. The power of modu- 
lating fluidly from one centre of tonality to another is the meaning of 

This fluidity is carried so far in twelve-tone music that key in its 
turn disappears. Here all twelve semitones of the chromatic scale are 
drawn upon equally and simultaneously. Like a temporary mode, a 
serial tone-row establishes the distribution of intervals, any departure 
from which reduces its distinctive character. Established centres of 
tonality are avoided; but the field of force is still the familiar tonal pull 
which is the fundamental acoustic reality of music. 

Mixed systems are both possible and desirable. Our key system 
incorporates two modes (major and minor) and accommodates others. 
Yet the historic transition from mode to key during the late renaissance 
was crucial: it marked the advent of modulation. 




There was one unavoidable exception to the diatonic character of 
the modes. This was the alternative of B natural (B durum, hard, from 
its supposed musical effect, or quadratum, square, from the use of the 
square form of the letter B to indicate it) and B flat (B molle, soft, or 
rotundum, round). 

The exception was unavoidable because of the acoustic realities. A 
diatonic scale includes the augmented fourth or tritonus (as F to B) and 
its enharmonic equivalent, the diminished fifth or quinta falsa (as B 
to F). These intervals have a complex ratio between the vibration fre- 
quencies of the tones composing them, and produce much discordant 
beating among their upper harmonics and combination tones. They 
have, therefore, an inherently restless and inconclusive effect which was 
highly distasteful to the medieval musicians, with their preference for 
the simplest ('perfect'} concords, namely the octave, fifth and fourth, 
and their only qualified acceptance even of the less simple ('imperfect') 
concords, namely the thirds and sixths, major and minor. Indeed, the 
medieval ear was more tolerant of plain discord than of the ambiguous 
augmented fourth and diminished fifth. 

To escape them, chromatic alteration of B to B flat was admitted. 
This note was not then regarded as chromatic; it was regarded as a 
slightly irregular extension of the diatonic material. But another method 
was to leave the B natural while taking the F sharp. This further exten- 
sion was evidently harder for the theorists to accept; and others 
followed. Their valid chromatic nature was still not fully recognised; 
but neither were they adopted, as B flat was adopted, into the material 
regarded as basically diatonic (musica verd). They were treated as a 
kind of falsification (musica falsa) or fiction (musica ficta). Once this 
concept of 'false music' or 'feigned music' was established, B flat also 
tended to be included within it. 

(77) Anon. II, Tractatus de discantu', Coussemaker, Scrlptorum, I, 

'Now false music was invented for two reasons, namely by reason of 
necessity and by reason of beauty of melody for its own sake/ 

The necessity was that of avoiding augmented fourths, diminished 
fifths, and imperfect or augmented octaves; the beauty was an ever- 
expanding conception, in practice largely connected with the growing 
disposition to sharpen leading notes at important closes. The theory of 
'false' or Teigned 5 music encouraged what would in any case have been 
the normal tendency of the period to leave chromaticisms unwritten 
for the performer to introduce at his own discretion. The terms musica 
falsa and musica ficta themselves, however, do not necessarily imply 



improvisation. They refer to chromatic alteration whether written in 
the text or introduced in performance. 


The means by which the diatonic material was explained by medieval, 
renaissance and early baroque theorists was not a seven-note scale 
recurring at the octave, but a six-note scale of three overlapping 
hexachords. Table I (p. 64) will show the structure. 

The point of the hexachordal system is that each hexachord contains 
only one semitone, and that one always in the same position: between 
mi and fa. 

In the 'hard' hexachord starting on G, mi is B and fa is C. No diffi- 
culty here: they are a semitone apart anyhow in the 'white note' 
diatonic scale. The B stays natural 

In the 'natural' hexachord starting on C, mi is E and/a is F. Again 
no difficulty; these two are also a semitone apart in the "white-note* 
diatonic scale. No B occurs in this hexachord. 

In the 'soft' hexachord starting on F, mi is A and fa is what? Not 
B, which would make it a tone instead of a semitone above mi; but 
B flat, regarded not as a chromatic alteration but as a different kind of 
diatonic B. 

The letters signifying these two forms of B were written approxi- 
mately in the shapes we have inherited as t) (square, i.e. natural B) and 
b (round, i.e., flat B). Originally notes in themselves, these symbols 
soon came into convenient use as signs for altering notes: at first only 
5s, but later, as genuine chromatic alteration grew, other notes as well. 
A further symbol approximating to our ft was never a note, but began 
as a sign cancelling a flat, and was subsequently extended to raising 
any note a semitone. The Iq became partly disused for some time in 
later medieval and renaissance music, ft being the normal cancellation 
of b and b of ft. 

It will be noticed that the lowest hexachord starts on G under the 
Greek style of Gamma (P), regarded as the theoretical bottom of the 
system; and that it ascends to e", regarded as the theoretical top. The 
system as a whole was called the gamut (a compression of gamma ut). 
Its limits were exceeded in practice. Beware of a rare use of b before/ 11 
or F without influence on the pitch, apparently as a mere warning that 
these are notes extra manum, outside the compass of the Guidonian 

To learn your gamut meant training yourself in singing or thinking 
of the hexachordal names so that you always knew where to expect the 
semitone from mi to fa. As you ascend the hard hexachord, you find 
its fourth note, Cfa, on a level with the first note of the natural 





1 S^ 
*2 O 3 


2.: ^ 

5 g - 
^ T? 
5^ 8 

^ o 






3 a 



Q I 

9 Q 





4. ^ 



H Q 





I 1 
s .? 

T? eq 

G P 
"* tU) C3 

.tS C '~ 



K 03 











hexachord, Cut (see Table I). The full name Is thus Cfaut; and if 
the continuation of the melodic line makes it convenient, you can think 
yourself over into that overlapping hexachord i.e. think of this C as 
its first note, ut 9 and carry on as before. This *thinking yourself over' 
from one hexachord into another is called mutation; and actually the 
standard note on which to mutate is one higher, i.e. D sol re. The 
standard mutation from a natural into a soft hexachord is G sol re ut. 
The return journey is similarly made. Reference to Table I will show 
that for a semitone between B and C you need to think In terms of a 
hard hexachord; for a semitone between E and F, in terms of a natural 
hexachord ; for a semitone between A and B flat, in terms of a soft 
hexachord. Thereafter musicaficta begins. 


TABLE II: Baroque Chromatic Signs (mainly 17th cent.) 

f b 
Flat: P P 

Sharp: ^ K f MX 

'Natural' (mainly to restore 5kj) b H 



TABLE III: Baroque Chromatic Signs (mainly 18th cent.) 
Flat: \) \> 

Sharp: ^ # X 

Natural (in modern usage) : fa & 


Double-sharp: ^ >^ M * X 



It remained the normal baroque method to cancel b by $ and $ by b- 
The difference is that whereas in modern notation b, I and iq give us 
fixed positions for their notes, in early notation b and $ give us semi- 
tone alterations down and up respectively, but not fixed positions. 
Thus D b may stand either for D flat, or for D natural restored after 
a previous D sharp; while D$ may stand either for D sharp, or for 
D natural restored after a previous D flat. Similarly in words: 'B sharp' 
is the normal seventeenth-century English term for what we call B 
natural. A 'flat key' meant a minor key; a 'sharp key' meant a major 
key. The importance of remembering these differences of notation and 
terminology is obvious. 

In the early baroque period we find k) at times used to cancel a flat 
against B, but not against other notes. 

(78) Galeazzo Sabbatini, Regola . . . per sonare sopra il Basso 
continue, Venice, 1628, Ch. XV: 

'In the case of the aforesaid letter [the note B] one should, it is true, in 
place off set fc) , which is its proper sign . . . but . . . # is commonly used/ 

But generally $ is not even mentioned. 

(79) Christopher Simpson, Compendium, London, 1665, ed. of 1732, 
p. 5 (alleds. have it): 

4 As for the b Flat . . . [except in the key signature] it serves only for 
that particular Note before which it is placed . . . that b takes away a 
Semitone from the Sound of the Note before which it is set, to make it 
more grave ex flat: This # doth add a semitone to the Note to make it 
more acute or sharp . . . [except in the key signature] it serves only for 
that particular Note before which it is applied.' 

(Leclair actually uses t) to restore a sharp present in the key-signa- 
ture: e.g. in A major, C natural is shown as Cb, and the subsequent 
restoration of C sharp is shown as C \\ ! He sometimes restores a B flat, 
present in the signature, with tj ! In none of this is he in the least 
consistent, though he does appear, fortunately, to be unique.) 

Not until the eighteenth century do we find our modern use of k) (as 
a general sign cancelling previous chromatic alteration by a semitone 
either down or up) becoming gradually standardised. 

The sensible /? for double flat never came into regular use. The 
modern x for double sharp was an early-eighteenth century improve- 
ment on >3, of which the horizontal stroke tends to be obscured when 
it stands on a line. Beware of confusion with x used for # by Penna, 
Leclair and some others, against figures of a figured bass; also with x 
used not as an accidental at all but as a frequent sign of ornamentation. 

A note already sharp or flat in the signature might be made double 
sharp or double flat by Jf or b- 



Return from double sharp or double flat to single sharp or single 
flat might be shown by &# and ift, by s, or (very commonly) not at all 

The use of, e.g., G for F double sharp and D for C double sharp is 
surprisingly common, not only in the seventeenth century, but also in 
the eighteenth (as in Johann Philipp Treibef s little Accurate Organist, 
Jena, 1704 where, since G sharp Is in the signature, F double sharp 
appears on the stave as \?Gl). 

During the middle of the eighteenth century our modern chromatic 
signs became increasingly standardised in all respects, the transition 
being virtually completed by the end of that century. 


The ordinary signature for a once-transposed mode (i.e. moved up 
bodily by a fourth) is one flat, and for the much rarer double trans- 
position (i.e. two fourths up, totalling one tone down), two flats; or 
the same results may be shown by accidentals. 

But the different parts of a polyphonic composition may be genuinely 
in different modes, and correctly bear different signatures (no flats 
against one flat, or even two), or show different accidentals. 

It will be appreciated that if the mode is not to be altered but merely 
transposed, C sharp becomes F sharp and B flat becomes E flat (for 
single transposition), etc. 


In baroque music, a key signature may often show one sharp or one 
flat short of the number required, the deficiency being subsequently 
made good by accidentals throughout the piece as required. This is a 
relic of modal notation, where what looks to us like a key signature 
was in fact a modal signature (i.e. showing transposition). 

Thus D minor, which we notate with a signature of one flat, grew 
out of the Dorian mode when its leading note, C, became habitually 
sharpened by the workings of musica ficta. The distinctively minor 
form of the scale has a flat sixth as well as a flat seventh; and it is this 
flat sixth which we notate as the B flat in our signature for D minor, 
adding the sharp leading-note (C#) and its concomitant the sharp 
sixth (5lq) by means of accidentals as required. But the B\\ is a dis- 
tinctive element in the Dorian mode, for which the signature therefore 
showed no sharps or flats. Once-transposed Dorian likewise led to G 
minor with a signature of one flat (B flat), the second (E flat) being 
accidentally supplied; etc. The habit lingered after the tonal facts had 
changed; and other signatures, by analogy, also appeared a sharp or a 
flat too few. We find, for example, Handel's Harmonious Blacksmith, 
which is in E major, with a. signature of three sharps instead of four. 



Since in such cases the accidentals needed to make up the number 
required for the key are ordinarily present, unless left out by oversight, 
no problem of interpretation arises here. 

. Another discrepancy affecting only the eye is the not uncommon 
habit of duplicating the sharps or flats of a key signature at the octave, 
where that octave lies within the stave. This may give the appearance, 
at first glance, of at least one sharp or one flat too many which is 
soon seen not to be the case. And where this duplication in the signature 
is not employed, a similar precaution may appear in the use of acci- 
dental sharps or flats against notes already shown thus sharpened or 
flattened in the signature, but at a different octave. None of this should 
be more than momentarily confusing. 


Genuine problems of interpretation over accidentals arise in two 
main aspects: the range of influence of written accidentals; and the 
necessity of supplying unwritten accidentals under the conventions of 
musicaficta. Though the two aspects are intimately connected, they can 
be considered most conveniently in separate chapters. 



Interpretating Written Accidentals 


The basic convention governing the influence of accidentals down to 
the end of the baroque period remained that given by Simpson, at (79) 
in Ch. IV, 4, above: each accidental affects only that note against 
which it is placed. 

Exceptions were extremely numerous, however, and were partly 
guided by subsidiary conventions. 


A convention which does not apply in early music is that governing 
accidentals strictly by the bar. The reason for this will readily be 

Bar-lines are found as occasional conveniences at very early periods, 
but first became frequent in organ scores and tablatures of the late 
sixteenth century, to make it easier to keep the place (i.e. when accom- 
panying unbarred polyphonic parts). They serve the same practical 
purpose of place-keeping in keyboard solos, where the notes themselves 
may often be written considerably out of vertical alignment; and the 
lutenists also employed them. But few of these early bar-lines occur at 
regular and consistent intervals of time; nor do they always appear at 
the same places in different copies of the same work. Except in dance 
or dance-influenced music, there were no regular intervals of time. 
There was no unit in the structure corresponding to our modern word 
"measure 1 in its sense of bar. 

In baroque music of the main period there was such a unit; but not 
until late in the eighteenth century did advantage begin to be taken of 
this change to build up our modern convention governing accidentals 
by the bar. We have to watch for cases such as Ex. 1, and conversely 
(but less frequently) Ex. 2. 

Ex. 1. Uncancelled accidentals within the bar, assumed to be 


Ex. 2. Accidental maintained across the bar-line: 

In Ex. 2, however, we should be more likely to find a sharp written 
not only against the second G but also against the third. 

By the middle of the eighteenth century, the force of an accidental 
was sometimes extended indefinitely until cancellation, the bar-line, 
however, still not being given a cancelling effect. 

(80) C. P. E. Bach, Essay, II, Berlin, 1762, 1, 38: 

'Unless an accidental [in the figuring] is cancelled, it remains in force.' 

The date at which the modern convention became fully established 
cannot be accurately stated; but there is no baroque music for which 
it was more than an incipient tendency, and perhaps no eighteenth- 
century music for which it is to be relied on unreservedly. For early 
music generally, it is a sound basis to ignore the bar-line when con- 
sidering the scope intended for a written accidental. 


(a) In the hexachordal system, an accidental B flat implies mutation 
into a soft hexachord, and it was therefore a regular convention that 
the influence of this flat should persist until further mutation was 
necessitated, either by a written B natural intervening, or by the melodic 
line moving beyond the limits of the hexachord (see Ch. IV, 3, above). 

By analogy, this convention also held good for flats other than B, 
but not for sharps, which have no place in the basic hexachordal 
system (though they were brought in by transposition, F sharp being 
mi, for example, where D is ut\ 

(b) As a loose consequence of this convention, we find in music at 
least down to the middle of the seventeenth century a general tendency, 
other things being equal, for flats to last but for sharps not to last. 


In theory, the basic convention that a written accidental affects only 
the note against which it is placed covered even the special case of the 
same note repeated twice or more times in succession, 

(81) Jacopo Peri, Euridice, Florence, 1600, Avvertimento: 

'[An accidental] is never to be introduced except on that note alone on 
which it is shown, even though there may be several repetitions of that 
same note.' 



In practice, great variability obtained. This is virtually admitted by- 
Peri in this very work, since he several times, when repeating the 
note in the bass, writes in a cancelling accidental which should not be 
necessary according to his own rule. We may infer that the mle might 
very easily be ignored by the performer in the absence of this precaution. 

When a note shown with an accidental is repeated, but the accidental 
is not repeated, we can (according to the sense of the music) either 
regard that accidental as cancelled or regard it as influencing the 
repetition or repetitions of the note (including repetitions which are not 
quite immediate, i.e. where one or two other notes of brief duration 

Where either a rest or the start of a new phrase intervenes, it is a 
virtual certainty that the accidental should be regarded as cancelled, 
except possibly in the case of flats where the compass of the hexachord 
has not been exceeded. 

When neither rest nor new phrase intervenes, the choice depends on 
our judgment of the musical situation ; but the tendency for the influence 
of the accidental to persist increased during the course of the seven- 
teenth century, and still more in course of the eighteenth. 


(a) A particular instance of repetition occurs in the ornamental 
resolution of discords. 

Ex. 3. Lodovico Grossi da Viadana, Concert!, Venice, 1602, IV, 
Taetare Hierusalenf, bar 3 (of bass): 

} _ M 

Cantusl ^ p- ppti^^ 

Cantus II .5 

Basso : 


The C sharp shown a third above and slightly to the left of the bass 
note A tells the accompanist to play an A major triad against the 
dissonant D of Cantus I (to sound the note of resolution against the 
note of dissonance in this way would not have been approved later in 
the baroque period, but was quite favoured both in practice and in 
theory at the date here in question). Hence the first C in Cantus I will 
have to be taken sharp as well as the second, though the accidental is 
only marked before the second. That accidental, in other words, exerts 
a retrospective as well as a prospective influence. 



For one case such as this which can be proved, there must be many 
hundreds which cannot be proved, but in which the same retrospective 
influence of a written accidental must, by analogy, be regarded as 
extremely probable. The probability is that most if not all cases of 
ornamental resolution written in this way, with an accidental marked 
against the main note of resolution but not against the previous orna- 
mental notes of resolution, are to be interpreted on the lines of the 
example above: Le. as if the previous ornamental note or notes of 
resolution were marked with the same accidental as the main note of 

(b) The retrospective use of accidentals is not, however, confined to 
ornamental resolutions. 

Ex. 4. Lodovico Grossi da Viadana, Concert!, Venice, 1602, II, 
Teccavi super numerum", bar 10 (of bass): 

Can tin 


Here the b marked before the first D of the cantus is a precautionary 
accidental, of a rather remarkable character. It warns us to take this 
D natural, and not, as the composer evidently expected that we should 
otherwise do, sharp. But why should we take it sharp? The only reason 
can be a retrospective influence exerted by the sharp marked against 
the second D. 

Attention has already been drawn by Knud Jeppesen (The Style of 
Palestrina, 2nd ed., London, 1946, p. 36) to a case of the same general 
character in Palestrina's eight-part mass, 'Confitebor tibi Domine', 
where a theme beginning with two repeated notes enters without acci- 
dentals, and elsewhere a tone higher with the accidental which this 
transposition necessitates shown before the second, only, of these 
repeated notes, though it must certainly influence them both in order 
to keep the theme unchanged. 

Ex. 5. Retrospective accidental in Palestrina. 

' i 1 I 

Ky- ri- e* e- Ky- ri- e e* 

The first of the two entries shown here, at Ex. 5 (a), is indistin- 
guishable, in its notation, from a chromatic step (F natural to F sharp); 



and it is only the lucky accident of its parallel a tone lower 

at Ex. 5 (b) which enables Jeppesen to prove both the first and 
the second F of Ex. 5 (a) are intended to be sharp. 

It is impossible to be sure how far this most tiresome irregularity 
extended in actual practice. Not very far, probably, in the particular 
form shown in this example from Palestrina; for even in an age when 
the performer was used to taking much of the responsibility for acci- 
dentals, the risk of misunderstanding would have been too great. He 
was evidently inclined, however, when he saw a written sharp, to come 
to the conclusion that sharps were in the air. He would not prolong a 
sharp as he would prolong a flat within the compass of the hexachord. 
But he might (or might not) allow its influence to pass a few notes 
backwards as well as forwards. 


See Ch. XXVII ff. But a warning may be given here that one of the 
most frequent misprints in early baroque editions is setting an acci- 
dental which should be above or below the bass note, i.e. as an indica- 
tion of the accompanying harmony, in front of that bass note so that 
it looks like (but is not) a chromatic modification in the bass line itself. 



Adding Unwritten Accidentals 


The convention that the performer should add unwritten accidentals 
where necessary, and might add them even where not necessary, is of 
importance for music down to and including the early seventeenth 

(82) Pietro Aron, Thoscanello de la Musica, Venice, 1523, p. 1 of 
adds, following Ch. 41 : 

'There are, among students of music, many arguments as to flats and 
sharps, that is whether composers should mark such accidentals ... or 
whether the singer is supposed to perceive and know the unknown secret 
of the places where such accidentals are required. I myself, who always 
was, and am, in favour of those who like to make their wishes clear . . . 
shall with all possible brevity treat the subject [unlike those who think 
that marked accidentals are for] beginners with no mind of their own/ 

(83) Steffano Vanneo, MS., 1531, tr. Vincento Rossetti, Recanetum 
de musica, Rome, 1533, III, 37: 

'The ears are considered the best interpreters, which can help you most, 
if you observe the parts of an accomplished singer, who when lie feels 
that he is producing a dissonant progression, at once little by little and so 
discretely, that it can scarcely be recognised and detected, either flattens 
or sharpens it, until a consonant and sweet progression strikes the ears/ 

(84) GiosefFe Zarlino, Istitutiont harmoniche, Venice, 1558, but cited 
from complete ed. of 1589, III, 252 (a contrary view): 

'There are some who in singing sharpen or flatten a melody in a case 
which the composer never intended as when they sing a tone instead of 
a semitone or the other way about, and other such things. By this they 
not only offend the ear but also commit innumerable errors. Singers 
should therefore take care to sing only what is written according to the 
mind of the composer/ 

(85) William Bathe, Brief e Introduction, London, c. 1590: 

'Thappendancy of the flat by the sharp, and of the sharp by the Flat is 
taken away, though by negligence and ignorance of prickers, we are oft 
driven to gather thappendencie by the course of the song/ 



(86) Thomas Morley, Plaine and Introduction, London, 1597, 
p. 88: 

"Because I thought it better flat then sharpe, I have set It flat. But if 
anie man like the other waic better, let him use his discretion." 

Theoretical recognition of musica ficta introduced more or less 
impromptu by performers was never as extensive as the practice of it; 
indeed it was at all times a controversial matter; but we find musical 
evidence for it well into the early seventeenth century, and its influence 
lingered indirectly in the general baroque attitude towards accidentals 
for some considerable period after it had become obsolete as a 


The most famous of ancient musical mnemonics is mi contra fa est 
diabolus in musica: mi against fa is the devil in music. 

B mi against Ffa ut gives a diminished fifth (quinta falsa) or an 
augmented fourth (tritonus). B mi against Bfa gives a diminished or 
augmented octave. 

(87) Pietro Aron, Thoscanello, Venice, 1523, p. I of adds, after 
Ch. 41: 

'The softening and tempering of the tritone [is] either by raising or 

(88) Charles Butler, Principles of Musik, London, 1636, p. 49: 

'But these harsh Discords, by the help of Flats and sharps, are reduced to 
their true Concords. For as the Tritonus [augmented fourth] either by 
flatting die sharp, or sharping the Flat, is made a true Diatessaron 
[fourth]; so the Semidiapente [diminished fifth], by the same means, is 
made a true Diapente [fifth]/ 

These intervals are not always, of course, to be eliminated by musica 
ficta; it is obvious that they were very frequently tolerated as elements 
in the harmony, and quite often as melody. But there are also a great 
many passages from which, at any rate as melody, they need to be 
eliminated. The standard corrections are as follows: 

Ex. 6. Imperfect fourths and fifths corrected: 

In each of these cases it will be seen that the note altered by musica 
ficta is the second of the two, which is obviously much easier to accom- 
plish impromptu than altering the first. 




Diminished and augmented octaves would in principle be regarded 
as in need of correction to perfect octaves by musica ficta. But there is 
one form of diminished octave which involves the simultaneous use of 
different hexachords in different parts, and which became a favoured 
idiom, particularly in England from Byrd to Purcell. 

In its milder, and presumably original shape, this idiom is a striking 
case of false relation, but produces no actually discordant harmony. 
In its more advanced and harsher shape, it is a clash of harmony, but 
produced by so natural a movement of its component melodies that its 
Tightness and logic are really unassailable. Our English taste for it 
throughout the greatest period of our musical history is an instance of 
the emotional depth and poignancy so typical of our national character 
at the time. It is, indeed, extremely beautiful. 

Ex, 7. The diminished octave clash: 

(a) melodic >\ }} harmonic 

Not only is it wrong to eliminate this form of mi contra fa by a misuse 
of musica ficta; it will normally be proper to introduce it even where 
not written, by taking the lower C (owing to its function as leading 
note) sharp even in the absence of a written accidental. Similarly in 
transposed positions: e.g. B flat in the upper part against B natural in 
the lower part. (See 6 below.) 

A somewhat comparable use of the augmented octave, quite typical 
although not amounting to so definite an idiom, may be seen in the 
following passage. The C sharp is marked, but even if not marked 
could have been introduced by musica ficta. The bass is correctly 
shown as C natural 

Ex. 8. Nicholas Carleton, 'Gloria Tibi Trinitas', bar 26, Mus. Brit. I 
(Mulliner Book), ed. Denis Stevens, London, 1951, p. 5: 

The first F in the bass (not the second) might optionally be taken 
sharp as an example of musica fata introduced not Tor necessity* but 
Tor beautification' : see 10 below. 




The minor third differs acoustically from the major third in giving 
rise to much more dissonant beating among its upper harmonics and 
combination tones. Moreover, any single note includes a major third 
fairly prominently among its upper harmonics, so that if a minor third 
is sounded on voices or instruments, it is always in faint but disturbing 
conflict with a major third on the same bass sounding of Its own accord 
among the overtones. 

The medieval ear was just as sensitive to this inherently restless and 
inconclusive character of the minor third as it was to the ambivalent 
character of the augmented fourth and the diminished fifth. The minor 
third was therefore not felt to be a tolerable element in any important 
close, least of all a final close, until well into the baroque period. The 
original solution was to exclude both major and minor thirds, leaving 
a bare octave or an octave with fifth. During the sixteenth century, 
however, thirds were increasingly used in final closes; but with the 
understanding that if the prevailing modality or tonality would have 
made them minor, they must be altered to major whether so written 
by means of an accidental or not (the so-called Tierce de Picardie or 
Picardy Third). This understanding persisted in some degree, though it 
is hard to be certain in what degree, throughout the seventeenth 

(89) Thomas Morley, flame and Easie Introduction, London, 1597, 
pp. 155-6: 

'[Morley's Master Gnorimus corrects, with an absence of comment 
more eloquent than words, his pupil's final close in the minor by writing 
in a Picardy Third.]' 

(90) Francesco Bianciardi, Breve regola^ Siena, 1607, ninth rule: 
In the final closes, one always ends with the major Third/ 

(91) Wolfgang Ebner, Germ, transl. Johann Andreas Herbst in his 
Arte prattica e poetica, Frankfurt, 1653, Rule 8: 

'But when it is a full cadence (vollkommliche Cadentia), the last note 
must always be taken with tke sharp/ 

(92) Lorenzo Penna, frimi Albori Musicali, Bologna, 1672, Bk. Ill, 
Ch. V, Second Rule: 

fi On the last note, a ... major third/ 

Penna's example to this rule shows that he has in mind a close which 
would otherwise be minor. 


Ex. SA. Penna, Picardy Third: 

- 5 ,7 

(93) Friedrich Erhard Niedt, Musicalische Handleitung, First Part, 
Hamburg, 1700, Ch. VIII, Rule 6: 

'[The final chord must be major regardless of what goes before, except 
that] French composers do the opposite, but everything is not good 
merely because it comes from France or has a French name/ 

Niedt's evidence is the first written statement of which I am aware 
casting doubt on the universality of the convention of the Picardy 
Third. As regards the French composers, Niedt's statement may well 
be too sweeping. As regards the general European practice, in spite of 
the categorical repetitions of the rule quoted above, it is hard to believe 
that the convention was so invariable as we are told. There are some 
French minor movements of the late seventeenth and early eighteenth 
centuries which have every appearance of being as well suited by a 
Picardy Third on their final closes as those of other nations. But on the 
other hand, there are very many minor movements of all nationalities 
at the same period which seem by no means well suited by a Picardy 
Third, but which are far more convincing musically when ending, as 
they have proceeded, in the minor mode. In practice I have very little 
doubt that they should be allowed to do so, and that whatever the 
nationality of the music, the Picardy Third should only be applied as 
late as this where the musical advantages of doing so are very evident. 

It may be mentioned that with the sole exception of Morley all the 
treatises quoted above, in this section, are dealing with figured bass. 
It is one matter to introduce a Picardy Third into the accompaniment 
of a close in which the composer has brought all the written melodic 
parts to a unison, as baroque composers so often do; it is another 
matter to introduce it into the last chord, for example, of a harpsichord 
solo, where this is written minor. Closes showing the Picardy Third by 
a written accidental are common in seventeenth century music of all 
kinds, and not scarce in eighteenth century music. But Picardy Thirds 
as a performer's option may well have become almost confined to 
figured bass accompaniments at latest by the beginning of the 
eighteenth century, and a matter for considerable discretion even 




(a) The principle of the leading-note depends on the fact that the 
semitone, being the smallest interval in Western use, draws the mind 
rather as an electric pole draws a spark when the opposite pole is 
brought near enough. As the word implies, we are led across this 
interval, and led more easily than across a larger interval This fact is 
among the realities which constitute the field of force we call tonality. 

(b) We find a very early tendency to make the third major when it 
is about to open out into a fifth but minor when it is about to close in 
to a unison; and to make a sixth major when it is about to open out 
into an octave but minor when it is about to fall to a fifth. 

(c) The same economy of effort lies behind the traditional injunction 
una nota super la Semper est canendumfa: a single note above la always 
must be sung fa, that is to say at the interval of a semitone. The rule 
means that where the melody (particularly at the summit of a phrase) 
rises from a note to one note (i.e. a 'single note') above it, and at once 
returns, that one note, whether so written or not, shall be at the less 
effortful interval of a semitone, and not at the more effortful interval 
of a tone. 

Reference to Table I (Ch. IV, 3) will show that where la is E la mi, 
the requisite semitone above it is hexachordally available as Ffaut; 
and that where la is A la mi re, the semitone is available as B (flat)/a; 
but that where la is D la sol re> there is no semitone fa available, so 
that an accidental E flat must be introduced. (Indeed in modern lan- 
guage, though not in hexachordal language, B (flat)/a is already an 

Thus E rising to F and at once falling again to E makes no problem: 
so it is written, and so performed. 

A to B and back to A requires the B to be taken flat whether so 
written or not. 

D to E and back to D requires the E to be taken flat whether so 
written or not. 

There will be exceptions in practice; and both in theory and in 
practice none of this applies in the Phrygian and Aeolian modes and 
their transpositions (the final notes B, E or A 9 or with a transposing 
flat in the signature, E, A or D, or with two transposing flats in the 
signature, A, D or G, suggest a Phrygian or Aeolian disposition see 
also Percy Scholes, Oxford Companion to Music, 9th ed., London, 
1955, s.v. 'Mode', for a succinct guide). 

(d) When the melody moves one note downwards and at once 
returns, there was again a certain tendency, but a far less clearly 
defined one, to make the interval a semitone, if not already so, by a 



written or unwritten use of musica ficta : thus D, C D in some contexts 
might become D, C sharp, D; but only where the C is being treated 
strictly as a leading-note (see 6 below) is there any reason to suppose 
that it must be sharpened in this way. 

Ex. 9. Una nota super la Semper est canendum fa: 


(a) By the leading-note is meant the note a semitone below (and 
pulling strongly towards) a tonic whether already established or newly 
reached. It is a fundamental element in tonality but not a prominent 
one in modality. Some modes have this semitone naturally, others can 
be given it by musica ficta. But the greater the insistence on a true 
leading-note, the greater the invasion of mode by key. The historical 
consequences of turning potential leading-notes (a tone distant) into 
actual ones (a semitone distant) by musica ficta were the advent of true 
modulation and the subordination of modality to key tonality. 

This historical revolution began late in the middle ages but was 
mainly the work of the renaissance. The convention which came to 
govern the sharpening of notes not so written, in order to make them 
into true leading-notes, depends on one of the most important structural 
concepts of those periods: the concept of the cadence. The primary 
cadence, or tenor cadence, was a fall of a tone from the note above the 
final or tonic to that final or tonic (cadence means fall 'that strain 
again; it hath a dying fall', wrote Shakespeare). The secondary cadences 
were the progressions of the other parts joined with the tenor cadence. 
A close was established when a tenor cadence and a descant cadence 
appeared together; the others (including the bass cadence which later 
became so prominent in defining key and key modulations) were 
optional additions. 

Ex. 10. Four-part 'close' with its 'cadences' : 

1 Descant 

2 Tenor 

3 Contratenor 

4 Bass 

GO 1 fill 

-jif H Q : n 


<g r f 

; s j " 




i pi 1 



It makes no difference In what inversions the cadences appear; if the 
tenor and descant cadences are present together In any parts, a 
has been defined. The normal procedure Is then to sharpen the lower 
note of the descant cadence so as to form a true leading-note, If (as in 
the Dorian, Mixolydian and Aeolian modes) it is not already such (i.e. 
as a modal subsemitonium). Where the necessary accidental is not 
marked, it should be supplied if the close is an important or final one, 
but not if the close is only taken in passing. 

(94) Thomas Morley, Plaine and Easie Introduction^ London, 1597, 
p. 94: 

'The base being a [descant] Cadence, the nature thereof requlreth a 

Ex. 11. Morley: 'descant cadence* in the bass: 

[p. 144:] 'In your third note you have a flat [descant] Cadence in your 
counter Tenor, which is a thing against nature, for everie Cadence is 
sharpe [with exceptions described as] passing closes and not of the nature 
of yours, which is a kind of full or final close . . .' 

(b) When the tenor cadence falls to A, its penultimate note B may 
be flattened accidentally to B flat, as an alternative to sharpening the 
penultimate note G of the descant cadence; and similarly in certain 
transpositions. This is the so-called Phrygian cadence (because it arises 
naturally in that mode, from which it should on no account be elimi- 
nated by musica ficta; the tenor cadence here being, exceptionally, a 
semitone fall which is sometimes called by a happy metaphor the 
'leading-note' of the Phrygian mode). 

Ex. 12. 'Phrygian cadence' : 

1 1 OH 

(95) Charles Butler, Principles ofMusik, London, 1636, p. 83: 

In true Cadences, the Binding half-note must ever be sharp. . . . 
Nevertheless the La Cadence is sometimes admitted; as in these examples. 5 

Ex. 13. Butler, 'Phrygian cadences': 

, Ji . i* -* J 

^j... " " | u o ti I u O r=^ 



(c) In spite of the crucial part played by modal theory and practice 
in the historical evolution of the true leading-note, it is not only in 
modal music that the structure of the close gives us the clue to the 
accidentals required. Something corresponding to the medieval and 
renaissance descant cadence is found in a majority of baroque closes. 
The 'Phrygian cadence* still accounts for a proportion of baroque 
half-closes; otherwise, the sharp or sharpened leading-note is the in- 
variable rule for every genuine close. It is the 'bass cadence* (bass rising 
a fourth or falling a fifth) rather than the 'tenor cadence' (fall of a tone 
to the tonic) which, in conjunction with the 'descant cadence* (suspen- 
sion resolved on the leading-note) now pre-eminently defines the close: 
but the treatment remains the same; and while the accidentals are less 
likely to be left unwritten, cases occur. The following from Monteverdi's 
own madrigal transcription of his monody, "Ariadne's Lament', is 
instructive in more ways than one. 

Ex. 14. Monteverdi, leading-notes not marked but requiring to be 
sharp (Malipiero's ed., VI, 16): 


In the first place, all the F's are meant sharp (not only is this musically 
obvious, but in a previous entry a tone lower they are all E naturals, 
no accidental being required). 

In the second place, bar two shows in effect a descant cadence (top 
line) with a tenor cadence (bottom line), and the leading-note E must 
be taken sharp although not written so (again musically obvious, but 
also proved, by the monodic version, Malipiero ed., XII, p. 165, where 
the corresponding passage has this leading-note marked sharp). We 
could have no clearer confirmation that accidentals have still to be 
supplied at need by the performer in the early seventeenth century. 

Later in the baroque period, as with other varieties of musicaficta, 
the need to supply sharps on leading-notes where they are not already 
present in writing rapidly diminishes, until we really find it remaining 
only in the special case of figured bass. 

(96) Agostino Agazzari, Del sonar e sopra * I basso, Siena, 1607, p. 6: 

'All cadences, whether intermediate or final, demand the major Third; 
some people, therefore, do not mark it, but as an added precaution, I 
recommend using the sign [when figuring basses].' 



(97) Francesco Bianclardi, Breve Siena, 1607, of the 
nine rules : 

'When [the bass] rises a fourth [or fails a fifth] we give it the major 
Third, and if it is not naturally major, it is made so by adding the Sharp, 
because it is by this progression that the close is established/ 

(98) Matthew Locke, Melothesia, London, 1673, Rule 3: 

S A Cadence is a Fall or Binding, wherein after the taking of a Discord or 
Discords, there is a meeting or Closure of Concords [on two bass notes] 
the last of which two Notes generally riseth Four, or falleth five Notes 
from the former; by which it is known (for the most part) to be a Cadence 
... the Thirds are Thirds Major, and so are to be Play'd on aH Bindings, 
and generally on all such bass Notes as the following Notes riseth four or 
falleth five Notes. 9 

The last statement is particularly to be noticed for its almost un- 
restricted scope. But we must also be prepared for passages in which 
the sharpening of the (apparent) leading-note by musicaficta is not 
intended, although the baroque criterion of a bass rising a fourth or 
falling a fifth is duly satisfied. 

(99) Adriano Banchieri, Dialogo Musicale, in 2nd ed. of UOrgano 
suonarino, Venice, 1611: 

'Here is an example of the freedom of the composer's mind in avoiding 
the accidental in leaps [by the bass] of a Fourth and Fifth.' 

Ex. 15. Banchieri: apparent cadences not treated as such (until 
the last) : 

It would be a gross misuse of musicaficta to make the first F sharp; 
though if the first note of the second bar had been G or B natural in place 
of B flat, it would have been an equally gross negligence not to do so. 

(d) It may be worth adding an early eighteenth century version of 
the formal cadences, to compare with the sixteenth century (and earlier) 
version given in Ex. 10 above. 

(100) Andreas Werckmeister, Harmonologia Musica, Frankfurt and 
Leipzig, 1702, p. 48: 

'These Cadences (Clausulae) take their name from the four principal 
parts, namely Treble, Alto, Tenor, and Bass; wherefore they are 
"Discaiitisirend", "Altisirend", "Tenorisirend", and "Bassirend'V 


Ex. 16. Werckmeister, formal cadences: 

Tniuiisirend Bassirend 

In a further example, Werckmeister shows the upper three cadences 
in different inversions, but retains the bass cadences in the bass part: 
it is hard to see why, since examples of both the descant cadence and 
the tenor cadence in the bass part are familiar enough in baroque 


The name 'double leading-note' has been appropriately suggested for 
the beautiful and characteristic formula of the late fourteenth and early 
fifteenth centuries, shown at Ex. 1?A below. 

But by the second half of the fifteenth century, the parallel perfect 
fourths which give this formula its special flavouring went out of 
fashion, and the penultimate note of the middle part was taken flat, 
as at Ex. 17B below. It is then no longer in double harness with the 
leading-note proper, but is a normal flat (or normally flattened) 

Ex. 17. (A) 'double leading-note'; (B) leading-note with flat sub- 

Where this formula is encountered in transposed versions, the same 
results, if not written, must be produced by musicaficta as required. 


The cadence named after Francesco Landini, the blind organist of 
early renaissance Florence, has the sixth degree of the scale inserted 
ornamentally between the leading-note and the final note, but is other- 
wise normal in its behaviour. 

Ex. 18. 'Landini Cadence': 

L it J- J JE 

-*-. n 


Where the seventh degree of the scale, in its function of leading-note, 
is sharp or has been made so by musicaficta, its neighbouring mediant, 



the sixth degree of the scale, either is or be so too, If It 

takes part In the cadence; and this it often does, where the resolution 
of the dissonance in the descant cadence is ornamental. The 
second occurs as a melodic interval in Monteverdi, but was not ordin- 
arily acceptable. Being so obvious, therefore, the necessary sharp on the 
sixth degree is very commonly left unwritten, even when the sharp on 
the seventh degree is present. It must then be supplied by musica ficta^ as 
in Ex. 19 below (in the original, the sharp sign in the bass stave indi- 
cating the major triad on the bass note E is misplaced against that E 
itself, by an obvious and indeed common misprint, instead of appearing 
in its proper place, the space above, where it stands, in the normal 
manner of Viadana's and some other very early Italian figured basses, 
for G sharp in the accompanying harmony). 

Ex. 19. Lodovico Grossi da Viadana, Concert^ Venice, 1602, II, 
'Peccavi super numerum', bar 19 (of bass): 


It was understood throughout the period of musica ficta that acci- 
dentals might be introduced not merely for necessity's sake (causa 
necessitatis) but also for beauty's sake (causa pulchritudinis). 

How widely this should be taken we do not know; but it seems highly 
probable that most of what a contemporary performer would have 
thought beautiful would have been covered by one or other of the 
specific conventions already considered in this chapter, and that there 
is little scope for further unspecified alterations of the accidentals. 

In applying the conventions, we should be much bolder than some 
modern editors have dared to be since the reaction against the reckless 
liberties taken a generation ago; not so bold, however, as the boldest 
(see Bibliography under Lowinsky). 

Outside the conventions, an occasional chromatic alteration may 
certainly suggest itself, which while difficult to bring under any par- 
ticular heading, has the stamp of authenticity and makes a pleasing 
improvement; but we have to be extremely careful not to spoil a good 
thing by insinuating an anachronistic flavouring which dilutes instead 
of enriching the original effect. (See Ex. 8 above, the? F sharp.) 

All in all, the main difficulty with musica ficta is in deciding how far 
to go. Some of the most successful of recent treatments of the problem 
will be found in Volumes I and IV of Musica Britannica, and in Helen 



Hewitt's edition, with its valuable prefatory remarks on this subject, 
of Odhekaton A . The reader is strongly recommended to compare the 

detailed solutions shown in these editions with the statements made in 
the present Chapter. 


There is a further line of investigation which when it has been 
thoroughly enough followed up should give us a clearer picture than 
the vague and inadequate contemporary theoretical evidence on 
musica ficta. This is by comparing the accidentals shown in different 
versions of the same piece, and in particular by lute and keyboard 
intabulations. A lute tablature which shows, not notes directly, but 
frets to be fingered, cannot be ambivalent (though it can be incorrect) 
as to the accidental intended. Keyboard tablatures showing sharps by 
loops or hooks attached to the letters are less unequivocal, but useful for 
comparison. In general, tablature books provide us with our most 
reliable information on accidentals known to have been introduced in 
actual performances; and enough concrete instances of this kind would 
yield a sounder theory than any of the textbooks provided the vague- 
ness of these textbooks being, perhaps, deliberate, but none the less 
tantalising to ourselves. 

I am indebted to Alan Fen-Taylor for the following typical example 
of this kind of comparison. The fact which emerges from such cases 
most clearly is the remarkable fluidity of the whole situation. The 
unreality of supposing that there can be any one correct and exclusive 
solution for the problem of accidentals, in any given piece of a date down 
to and including the early baroque period, could not be better shown. 

Ex. 19A. Cipriano de Rore (1516-65), (a) madrigal 'Ancor che col 
partire', with keyboard intabulations: (b) by Antonio de Cabezon, 
posthumously published in Obras de musicapara tecla, arpay vihuela, 
Madrid, 1578; (c) by Bernhard Schmid, Zwey Bucher, Strassburg, 
1577; (d) by Andrea Gabrieli, Terzo libra de Ricercari, Venice, 1596. 


0*' t , T 


. * MM^^ + 





The Place of Embellishment 

Music hardly ever, if at all, consists only of its basic progressions. It 
is embellished. 

The embellishment is everything in the music which can be changed 
without affecting the basic progressions. It is partly a matter of 
rhythmic and harmonic detail, but mainly of melodic outline. It is, in 
fact, what we normally call the figuration. 

We are so accustomed in the modern West to the composer providing 
his own figuration in its entirety that we may not realise how different 
the situation has been at other periods. There have been whole schools 
depending on spontaneous invention within a traditional framework. 
Even when formal composition became more important than impro- 
visation, it continued to benefit from the same source; the improvisation 
of the performer continued to pour over into the resources of the 
composer. Our present instrumental music largely derives from the 
spontaneous adaptation of vocal and dance material to the idiomatic 
requirements of renaissance organists, lutenists, viol-players and others. 
Renaissance and baroque singers made a crucial contribution to florid 
vocal music. Even in the nineteenth century, composer-pianists like 
Chopin and Liszt were still enriching keyboard figuration at their 
finger ends; but this time-honoured flow of impromptu invention from 
performer to composer (whether in the same or different persons) must 
have thereafter largely dried up (Lukas Foss and others are re-inventing 

It is the wealth of passing detail, the felicity, unexpectedness and 
exuberance of the figuration which makes music out of its mere basic 
progressions. It may not ultimately matter where the figuration comes 
from, but it matters very much that it should be present. Our problem 
here is that even in the baroque period composers still left substantial 



proportions of their figuration to be supplied, more or less 
by the performer. Hardly any modern performers are yet of 

doing this. The main responsibility thus passes to the editor, who can 
at least train himself to provide a working version in writing. This is 
not an ideal solution, but it is greatly preferable to leaving the figuration 
in its incomplete state as originally composed and published. 


(101) Desiderius Erasmus, note on Corinthians, XIX, 19, transl. J. A. 

Froude, Life and Letters of Erasmus, London, 1894, p. 116; 

'Modern church music is so constructed that the congregation cannot 
hear one distinct word.' 

(102) Archbishop Cranmer, letter to Henry VIII, 1544: 

'In my opinion, the song that shall be made [as music to the Litany], 
would not be full of Notes, but as near as may be, for every syllable a 
note, so that it may be sung distinctly and devoutly.' 

(103) Queen Elizabeth, 49th Injunction, of 1559: 

'. . . a modest distinct songe [for the body of die service] that the same 
may be as playnely understanded, as yf it were read without syngyng 
[but for the anthem] the best sort of melodic and music that may be 
conveniently devised, having respect that the sentence of the Hymne may 
be understanded and perceyved.' 

(104) Pope Marcellus, address to the papal choir, 1555, cited Grove's 
Dictionary, 5th ed., London, 1954, s.v. Talestrina*: 

'[Church music must be sung] in a suitable manner, with properly 
modulated voices, so that everything could be both heard and properly 

(105) Charles Butler, Principles of Musik, London, 1636, p. 116: 

'Too much quaint Division, too much shaking [trilling] and quavering 
[breaking down into rapid variations: cf Cotgrave's Dto., London, 1650, 
s.v. "quaver"] of the Notes, all harsh straining of the Voices beyond their 
natural pitch, as they are odious and offensive to the ear; so do they drown 
the right sound of the words . . / 

(106) Giulio Caccini, Nuove Musiche, Florence, 1602, Preface, transl? 
Playford (from 2nd ed. of 1607), Introduction to the Skill of Musick, 
London, 1654 (1664 on), cited from ed. of 1674, p. 37: 

'Hitherto I have not put forth to the view of the World, those Fruits of 
my Musick Studies employed about that Noble manner of Singing, 
which I learnt of my Master the famous Scipione del Palla in Italy; nor my 


Compositions of Ayres, Composed by me, which I saw frequently 
practised by the most famous Singers in Italy, both Men and Women: 
But seeing many of them go about maimed and spoil' d, and that those 
long winding Points were ill performed, I therefore devis'd to avoid that 
old manner of running Division which has been hitherto used, being 
indeed more proper for Wind and Stringed Instruments than for the 
Voice. And seeing that there is made now a days an indifferent and con- 
fused use of those excellent Graces and Ornaments to the good manner 
of Singing, which we call Trills, Grupps, Exclamations of Increasing and 
Abating of the Voice, of which I do intend in this my Discourse to leave 
some footprints, that others may attain to this excellent manner of 
Singing: To which manner I have framed my last Ay res for one Voice to 
the Theorbo, not following that old way of Composition whose Musick not 
suffering the Words to be understood by the Hearers, for the multitude of 
Divisions made upon short and long Syllables, though by the Vulgar 
such Singers are cryed up for famous. But I have endeavoured in those 
my late Compositions, to bring in a kind of Musick, by which men 
might as it were Talk in Harmony, using in that kind of Singing a certain 
noble neglect of the Song (as I have often heard at Florence by the Actors 
in their Singing Opera's) in which I have endeavoured the Imitation of 
the Conceit of the Words, seeking out the Cords more or less passionate, 
according to the meaning of them, having concealed in them so much as 
I could the Art of Descant, and paused or stayed the Consonances or Cords 
upon long Syllables, avoiding the short, and observing the same Rule in 
making the passages of Division by some few Quavers to Notes and to 
Cadences, not exceeding the value of a quarter or half a Semibreve at most. 
But, as I said before, those long windings and turnings of the Voice are 
ill used, for I have observed that Divisions have been invented, not because 
they are necessary unto a good fashion of Singing, but rather for a certain 
tickling of the ears of those who do not well understand what it is to sing 
Passionately; for if they did undoubtedly Divisions would have been 
abhorr'd, there being nothing more contrary to Passion than they are; 
yet in some kind of Musick less Passionate or Affectuous, and upon long 
Syllables, not short, and in final Cadences, some short Points of Division 
may be used, but not at all adventures, but upon the practice of the 
Descant . . , certain it is, that an Ayre composed in this manner upon the 
conceit of the words, by one that hath a good fashion of singing, will 
work a better effect and delight, more than another made with all the 
Art of Descant, where the Humour or Conceit of the words is not 

(107) Chr. W. Gluck, dedication to Alceste, Vienna, 1769: 

'When I undertook to compose Alceste, I set out to divest the music 
entirely of all those abuses with which the false vanity of singers, or the 
too great complacency of composers, has so long distinguished the Italian 
opera. ... It was my intention to confine music to its true function of 



serving the poetry in the interests of the expression and of the 

story; without interrupting the action, or stifling it with and 

superfluous ornaments . . . my first and chief care, as a com- 

poser was to aim at a noble simplicity; and I have accordingly shunned 
all parade of unnatural difficulty, in favour of clearness/ 


(108) Benigne de Bacilly, UArt de bien chanter, Paris, 1668, p. 135: 

'A piece of music can be beautiful and please not, for want of being 
performed with the necessary embellishments, of which embellishments 
the most part are not marked at all on paper, whether because in fact they 
cannot be marked for lack of signs for that purpose, or whether it has been 
considered that too many marks encumber and take away the clearness 
of a melody, and would bring a kind of confusion; besides, it is useless to 
mark things, if you do not know how to fashion them with the appro- 
priate refinements (former avec les circonstances necessaries), which 
makes all the difficulty/ 

(109) Anthony Aston, Brief Supplement to Colky Cibber, Esq., 
1748, ed. R. W. Lowe in Gibber's Apology, II, London, 1889, p. 312: 

'[The boy singer Jemmy Brown] when practicing a Song set by Mr. 
Purcell, some of the Music told him to grace and run a Division in such a 
Place. O let him alone, said Mr. Purcell; he will grace it more naturally than 
you, or I, can teach him! 

(110) Gentleman's Journal, London, Nov. 1692: 

'[Counter-tenor solo in PurcelTs "Ode on St. Cecilia's day"] was sung 
with incredible Graces by Mr. Purcell himself [actually by Pate]/ 

(111) Roger North, Brit. Mus. MS. Add. 32533, English early 18th 
cent.,/. 106 v.: 

It is the hardest task that Can be to pen the Manner of artifical Gracing 
an upper part; It hath bin attempted, and in print, but with Woefdl 
Effect . . . the Spirit of that art is Incommunicable by wrighting, therefore 
it is almost Inexcusable to attempt it.' 

(112) Johann Adolf Scheibe, Critische Musicus, I, 12, Hamburg, 

'To a good method belong the appoggiaturas, the grace-notes, the 
trills, the alteration or elaboration of the notes, certain small, agreeable 
additions and variations, and many other things which are better listened 
to than described. 

[I, 6 :] 'All embellishments, all little graces, and all that is understood by 
the method of playing, he [J. S. Bach] expresses in [actual] notes, and not 



only deprives his pieces of beauty and harmony but makes the melodic 
line utterly unclear.' 

(113) J. A. Bimbaum in reply, repr. in Mizler's Musikalische Biblio- 

thek, Leipzig, April, 1738, I, iv: 

4 {J. S. Bach] is neither the first nor the only man to write thus .... It is 
certain that what is called the method of singing or playing is almost 
universally valued and thought desirable. It is further beyond doubt that 
this method can content the ear only when introduced in suitable places 
and that it is bound, on the other hand, to distress the ear markedly and 
destroy the fundamental melody if the performer introduces it at the 
wrong moment. Now we know from experience that its introduction is 
ordinarily left to the free choice of singers and players. If all such people 
were well enough brought up in all that is genuinely beautiful in the 
method; if they always knew how to introduce it where it may function 
as a genuine embellishment and special emphasisation of the fundamental 
melody; then there would be no point in the composer writing down in 
notes what they knew already. But only a minority have enough know- 
ledge, and the remainder destroy the fundamental melody by an unsuitable 
introduction of the method, and often actually introduce embellishments 
of a kind which could readily be ascribed by those unacquainted with the 
truth of the matter to a mistake by the composer. Thus every composer 
including [J. S. Bach], has the right to bring the erring back into the true 
way by indicating a proper method in accordance with his desires, 
thereby seeing to the safeguarding of his own reputation.' 

(114) Louis Bollioud-Mermet, De la corruption, Lyon, 1746: 

4 [All sonatas sound alike now, being overladen with the same embel- 
lishments.] One could easily persuade oneself that he who plays music in 
such a complicated and unnatural manner is an evil-doer on whom this 

labour was imposed for punishment.' 

(115) [John Hawkins] on Agostino Steffani, Memoirs of the Life of 
AS. [? London, c. 1740], p. ii: 

*. . . the exactness he required in the performance of his music, which 
was so remarkably great, that he would never admit of any divisions, or 
graces, even on the most plain and simple passages, except what he wrote 
himself: nor would he, with regard to his duets in particular, even suffer 
them to be performed by any of those luxuriant singers, who had not 
sense enough to see the folly of sacrificing to the idle vanity of displaying 
their extent, or power, of voice, not merely the air, but frequently the 
very harmony of an author's compositions.' 

(116) [L'Abbe Laugier] Apologie, [Paris], 1754, p. 71: 



'There ought to be a law preventing all singers and all 
up the orchestra from changing any tiling in the melody of which the 
character is traced out for them, with orders to restrict themselves scrupu- 
lously to the notation before their eyes. There ought to be a like law 
obliging all masters who give instruction to instill in their pupils the habit 
of literal performance.' 

(117) G. C. Weitzler, in Marpurg's Historisch-Krlfische Beytrage* 
Berlin, III, 2, 1, 1756: 

'With regard to the embellishments, I think it advisable to write them 
out in notes only when they are ordinary quick passages; with others 
which come partly from the natural facility of the fingers, partly from 
luxuriant invention, it is useless to do so. Either they cannot be indicated 
in notes, or else good taste will reject such embellishments so soon as one 
realises that they are supposed to be these and none others, thus and not 
otherwise at a given place; in short, as soon as one misses the fortuitousness 
so essential to them. [Such written-out embellishments do not sound 
natural; notation cannot always ensure a proper performance*] A musical 
person with good interpretative powers will never play in the same way 
but will always make modifications [in the notes] in accordance with the 
state of his feelings.' 

Marpurg follows this letter with a biting and satirical attack on 
Weitzler for his championship of the 'fortuitous' element in musical 
inspiration. The word evidently exasperated him and he does less than 
justice to Weitzler's real argument. Even so he adds: 

(118) Marpurg, loc. at., 1756: 

'It is true that the larger embellishments ... by which one modifies 
the repeat of a piece, are arbitrary. But are not some always better than 
others? . . . Moreover, must not these larger embellishments always 
reflect the content and emotions of a piece? If an ignorant keyboard 
player not gifted with judgement of feeling plunges fortuitously into an 
Adagio with embellishments better suited to an Allegro, will this embel- 
lishment be natural? The ear of a connoisseur will badly miss the per- 
former's skill in such a case.' 

Weitzler's reply (p. 357) was dignified. He had not wanted to over- 
throw the rules for good embellishment; but there are some kinds 
which cannot be satisfactorily written out in notes, because they should 
be spontaneous. A good teacher will play them in the right places, and 
a good pupil unconsciously pick them up; they sound fresher and 
express the personality of the performer better if they are his own 
invention. And so on at considerable length. His arguments were un- 
doubtedly of a kind which many, perhaps most of his contemporaries 
would still have endorsed; even his immediate opponent in this wordy 



duel, Marpurg, shared with him the fundamental assumption that 
decorative embellishment was required and that much of it was best 
supplied by the performer provided it was not done 'fortuitously' ! 

(119) F. W. Marpurg, Anleitung, Berlin, 1755, 2nd ed. 1765, 1, 9, p. 36 : 

It has been found that it is not enough to play off a number of pre- 
scribed notes according to their succession and time-values. It has been 
found necessary to render this succession agreeable to the ear, and to rid 
the composition of any roughness still adhering to it. From this have 

originated the performer's embellishments . . .' 

(120) Dr. Charles Burney, History, London, 1776, ed. of 1935, 

London, II, 443 : 

'Purcell, who composed for ignorant and clumsy performers [!] was 
obliged to write down all the fashionable graces and embellishments of 
the times, on which accounts, his Music soon became obsolete and old 
fashioned; whereas the plainness and simplicity of Corelli have given 
longevity to his works, which can always be modernised by a judicious 
performer, with very few changes or embellishments.' 

(121) Dr. Burney, in Rees' Cyclopaedia, London, 1819, s.v. 'Adagio': 

'An adagio in a song or solo is, generally, little more than an outline 
left to die performers abilities to colour ... if not highly embellished, 
[slow notes] soon excite languor and disgust in the hearers.' 


When Scheibe, at (112) above, criticised J. S. Bach for writing out 
so much of the ornamental figuration which almost any other composer 
of the time would have left to be supplied more or less impromptu by 
the performer, he had in mind that a performer who sees only the main 
notes (chiefly the harmony notes) of the melody written down tends to 
give these main notes their natural weight and to festoon his own 
ornamental figuration in between with the requisite lightness and 
elegance. This makes everything clear and in proper balance. But if the 
ornamental figuration has been fully written out by the composer, it 
all looks equally important to the eye, and the performer tends to play 
or sing the ornamental notes as emphatically as the structural notes. 
He weights them too regularly and measures them too literally. The 
melody is indeed obscured, as Scheibe wrote, and the effect is very 
ponderous and unsatisfactory. Cf. also (108) above. 

Bach's departure from baroque custom has the advantage of enabling 
us to learn from the Courante avec deux Doubles (First English Suite) 
or the Sarabands with Les Agremens de la mSme Sarabande (Second 
and Third English Suites) how he varied his own figuration; from his 
Vivaldi transcriptions, how he varied another composer's figuration; 



from the slow movements of his Violin Sonatas, how we should em- 
bellish such baroque slow movements when they survive only in the 
customary bare outlines. Handel's second slow movement to his A 
major Violin Sonata, which he altogether exceptionally wrote out in 
full, has quite the sound of a Bach slow movement, and it is obvious 
that this style was far more general than we used to suppose, though it 
was not generally committed to writing. But it is also necessary to 
realise that being essentially of the nature of improvisation, it must be 
made to sound impromptu even where it is in fact prepared beforehand 
by the composer, the performer, the editor, or a combination of these. 
How much of it is written, memorised or improvised is not important, 
provided that it has a genuine spontaneity in performance. As so often 
in baroque music, it is the freshness and incisiveness of the effect which 
really matters. Heaviness is the real danger; and this applies not only 
to ornamental figuration in slow movements, where it was so often left 
to the performer, but also in quick movements, where it was normally 
written out in adequate quantity, and impromptu variations, though 
usual, were in no way obligatory. 

(122) James Grassineau, A Musical Dictionary, London, 1740, s.v. 

'Canto Fioretti, is a song full of diminutions, graces, passages, etc., and 
is indeed figurate counterpoint.' 

(123) Jean- Jacques Rousseau, Dictionnaire de Musique, Paris, 1768, 
s.v. 'Passage': 

'[Passage-work is music] composed of many notes or divisions which 
are sung or played very lightly/ 
I can hardly emphasise enough the importance of those last two words. 


The disastrous effect on a slow movement by Corelli, Vivaldi, Handel, 
in some cases Purcell, etc., of failing to supply any ornamental figura- 
tion where only the bare framework has been provided by the composer 
in writing may be measured by reducing a familiar movement of 1801 
to the notation in which it might have survived had it been composed 
half a century earlier (solution at the end of this chapter). 


Violin etc. 

Realisation of b.c. by some modern editoi 

Basso Contini 





'Ornamentation' can conveniently be kept for the longer, free em- 
bellishments, and 'ornaments' for the shorter, specific embellishments 
such as trill, appoggiatura, etc. They were not clearly distinguished at 

the time, but List A below tends to imply ornamentation, and List B 

Terms for Embellishment 







List A 




agremens or 
agrements (du chant, de 



Ex.21. Beethoven: Spring Sonata, Op. 24, MovL I, as actually 





The sixteenth-century art of ornamentation, which was inherited from 
a much older tradition, was used on a variety of material. This included 
many kinds of canto fermo (plainchant, chorales and other hymns, 
popular melodies, etc.); ground-basses (traditional or taken from 
existing bass-parts or specially composed); and complete polyphonic 
works (mainly motets, chansons and madrigals) of which each part in 
turn, or even several together, might be involved in the ornamentation. 

Each of these practices led incidentally towards new forms of com- 
position, some of them of the greatest value and importance, such as 
the early baroque English music for harpsichord and organ (the 
'virginalists') or the Lutheran German choral prelude. 

The most characteristic common feature of sixteenth-century orna- 
mentation is that it affects the melody but not the harmony, with the 
result that it can be, and frequently was, performed simultaneously 
with the original. 


Treatises such as Ganassfs Fontegara (Venice, 1535) give written 
instructions, with examples, on introducing more or less improvised 
ornamentation such as had previously been given by word of mouth. 

(124) Adrianus Petit Coclico, Compendium musices, Niirnberg, 1552 
(of his first example): 

Ex. 2U. 

Cantus. simplex elegans 

'This is the first embellishment which Josquin taught Ms own pupils/ 

The examples are specimen runs, turns and similar small figurations 
filling in the intervals of notes at various intervals apart. The manner 
in which these figurations were combined into music can already be 
demonstrated for the fifteenth century. The type of Venetian song 
called Justiniana after its originator and chief exponent, Leonardo 
Giustiniani (c. 1383-1446), depended on the florid ornamentation of its 



upper melody in performance. W. H. Rubsamen (Acta Musicologica, 
XXIX, 4, Basel, 1957) has been able to compare one of these with a 
performing version as taken down from improvisation. He prints both 
versions complete; the opening is reproduced here by his kind permis- 
sion and that of Barenreiter-Verlag, Kassel and Basel. 

Ex. 22. 15th-century ornamentation: 


Aime sosptri 

PetruccL, Frottole L&. 

II" 7 ' j 


Ay - 

J. J " = 
pi - rC 

^ft_ 1 




- me 

sos ------ 

I" .: I f fzz 

.^ ^ -r 

tt. t J J 

J^'g j = 

.1 n.^- 


This Is solo music, accompanied with a lute (which be played 

by the singer) or viols. The ornamentation Is therefore performed by 
Itself, without the plain original. 


By the middle of the sixteenth century, we have much fuller treatises 
on Improvisation, both in extemporising new compositions and In 
ornamenting existing ones. 

(125) Diego Ortiz, Tratado de Glosas Sobre Cldusulas, Rome, 1553, 

'[Advice to viol players:] The first and most perfect manner consists of 
when one wants to pass from the passage or embellishment on any note 
found in the original to another succeeding note, the last note of the 
embellishment being the same as the note which was embellished, as these 
examples show. 



'This is the most perfect way as the embellishment begins and ends on 
the same note, making the leap in the same way as the plain melody, so 
that it can have no imperfections [i.e. forbidden consecutives]. 

'The second manner is freer, since, at the moment of progressing from 
one note to the next, it does not move like the plain notes, but on the 
contrary as these examples show.' 

Ex. 24-25: 

'This manner is necessary because with its freedom many pretty fancies 
and pleasant preluding can be made which cannot be made with, the first 
alone; therefore I supply it in some parts of this book. It may be thought a 
fault that at the movement from one fourth note (qu arto punto) to another, 
since it does not make the same leap as in the first manner, the other 
parts may form two perfect consonances [forbidden consecutives] with 
it; but this is unimportant since at that speed they cannot be heard. 

'The third manner is: to deviate from the composition, going by ear 
without being certain of what you are playing. This is the habit of some 
who want to demonstrate whatever skill they have, but depart from the 
composition at the wrong time and without rhythm, landing up eventu- 
ally on a cadence or on notes known to them. This is an objectionable 
process because it does not agree with the composition and can therefore 
have no perfection. 

'[Part II gives three methods for a viol and a harpsichord.] First ... I 



cannot show the fantasy since everybody plays it in his own style; but 
... on the harpsichord it should consist in well-ordered harmonies; the 
viol enters with some graceful passages, and when it sustains some plain 
notes, the harpsichord should answer it [with similar passages] at the 
right time; you should do some fugues, the one waiting for the other in 
the way in which you sing concerted [impromptu] counterpoint. 

4 [The second way of Part II is for the viol to improvise an embellished 
version of a given plainchant; the harpsichord meanwhile playing the 
original version but with a simple, mainly chordal, accompaniment 
improvised above it: i.e. "with accompanying chords and some counter- 
point fitting in with the Ricercada which the viol is playing 7 '. The same 
is also given for two viols alone without harpsichord accompaniment, 
when "the counterpoint will remain as perfect as if it had been written 
for one part alone" in addition to the plainchant.] 

'[The third is to:] Take a Madrigal or a Motet or any other piece and 
adapt it to the harpsichord in the usual way; the viol player can now play 
two, three or more variations over each given piece. When the soprano 
part is the one which is being embellished it is more pleasing when the 
harpsichord player does not play the soprano [i.e. its plain version at the 
same time as the embellished version, though otherwise Ortiz treats it as 
standard to sound the plain part and the embellished part at the same 
time]. . . . 

f He who uses this [third] manner should note that it is different from 
what we treat in the first book, namely to play together on four or five 
viols; there, it is necessary for good effect to suit the counterpoint [i.e. the 
embellishment] to the part [i.e. the particular line of the original poly- 
phonic composition] being played, holding on firmly to it . . . but in 
this manner here, it is not necessary to follow one and the same part; for 
though the bass must be the main argument, yet the player may leave it 
and pass over into the tenor, alto or treble as he pleases, taking the most 
suitable of each part. The reason is that the harpsichord plays the work 
completely in all its parts; the viol accompanies and gives grace to what 
the harpsichord plays, the changing sound of the strings giving pleasure 
to the listener. 5 

The many beautiful examples of this style which follow are accessible 
in the modern reprint, for which see my Bibliography. 

(126) Hermann Finck, Practica Musica, Wittenberg, 1556, Lib. V, 
[p. 8]: 

'[Advice to singers:] Truly in my opinion embellishments both can and 
ought to be scattered through all the voices [parts], but not all the time, 
and indeed in appropriate places, and not simultaneously in all voices, but 
let them be embellished in a fitting situation, remaining in their own 
places, so that one embellishment can be heard and picked out expressly 
and distinctly from another, yet with the composition whole and 



(127) Lodovico Zacconi, Prattica dl Venice, 1592, LXVI, 

p. 58: 

* Nimble ininds are continually discovering new embellishments . . . 
even to works previously embellished . . . like so many well-trained 
birds ... 

'[Must be taught by ear] because it is not possible to give in an example 
the exact measure for performing them ... it must be done to perfection, 
or left alone . . . 

'Perfection and beauty lie in the measure and tempo . . . tMs is the 
greatest difficulty in ornamentation, requiring more care and study than 
the ambition to make long strings of figures together. Indeed, the singer 
who with a little ornamentation in good time goes not too far afield 
will always be better appreciated than another who digresses far too much, 
whether in good time or not . . . 

"At the start of a [polyphonic] vocal composition, where the other 
voices are silent, you should not begin with an embellishment, nor 
immediately after the start . . . the beauty of an embellishment arises 
from the enchanting and graceful movement of the parts when one of 
them moves faster [than the other: i.e. from their mutual contrast]. Thus 
opening passages unless they are in [homophonic] part-writing should 
always be set forth simply and clearly so that the entry of each part may 
be the better heard. 

'[Moreover, until you know what the other singers are keeping in 
reserve, you should go cautiously] and then gradually unfold your own 
ornamentation [yet do not] save up everything for the end, having left 
the middle dry and empty. 

'[Crochets are too short to be generally suitable for ornamentation. 
Minims may be treated more freely, and still more semibreves and breves] 
but care should invariably be taken not to obscure the words. 9 


The 'unaccompanied' polyphony of the sixteenth century has been 
such a bye-word for purity that it comes as something of a surprise to 
learn that it was not only quite often accompanied by instruments, but 
actually embellished in performance with ornamental figuration, even 
when sung unaccompanied. This does not mean that unaccompanied 
and unornamented Josquin or Palestrina is unhistoric: on the contrary, 
all these methods of performance are authentic, and the choice rests 
with the performers. 

Where there are more performers than one to a part, either there was 
careful rehearsal to secure harmonious results, or the ornamentation 
was left to soloists with the others holding the plain original. A high 
degree of passing dissonance and a certain degree of 'incorrect' pro- 
gression was tolerated, because disguised by the speed at which the 



ornamentation proceeded. It was for the ear to decide how much. We 
must remember that the finest choirs were in very regular practice 
indeed, and that they were schooled to the last subtleties of disciplined 

In modern performance, it would be difficult to recover the art of 
improvising this kind of ornamentation, even where the choir Itself 
approaches sixteenth-century standards, because the training in extem- 
porising counterpoint at sight which was normal at the time, and from 
which the immense facility required for such improvisation was derived, 
is no longer given. It would be possible for a modem musician to 
acquire it from treatises such as Morley's Plaine and Easie Introduction 
(London, 1597); but without doing that, he may wish to make some 
experiments in writing out ornamentations and performing them. He 
should then work from the models of sixteenth and early seventeenth 
century polyphonic ornamentation which survive. See also Appendix III. 

Two shorter illustrations may be given here. The first shows a selec- 
tion of the figures suggested by Ortiz for ornamenting a given cadence. 
The second is a combination according to his directions of the opening 
of two of his full workings, one in the treble (not accompanied by the 
plain part) and the other in the bass (accompanied by the plain part). 
Viols were the medium primarily intended here; but 'brass' (cornetts, 
trombones), wood-wind (recorders, crumhorns) or voices (to the words 
of the original chanson) would have been, contemporary alternatives. 
Readers interested in doing so can readily complete the piece from the 
modern reprint of Ortiz (see Bibliography). 

Ex. 26. Ortiz, specimen figurations : 

/-N 1, 









| || | J 



1 j -fl 

.?.^ ,. 5 


Ex. 27. Ortiz, ornamentation of chanson, *Q felice mei' : 



Early Baroque Ornamentation 


Renaissance methods of ornamentation were carried into early baroque 
music with very little change other than the general disposition to 
restraint due to the expressive ideals of the new Italian monodists. 
The effect of the ornamentation remained essentially melodic. 

It is the long appoggiatura which gives later baroque embellishment 
its special character. This had its source in the suspensions of sixteenth- 
century harmony; and its beginnings can be traced at least as far back 
as Ortiz. But they were very small beginnings, and they were not much 
bigger in early baroque music, during the first decades of the seventeenth 
century. They were not yet big enough to produce essential changes in 
the harmony. 

On the other hand, there was no intention in the new monody, at 
any rate, of sounding the plain original simultaneously with its orna- 
mentation. There was no harmonic obstacle to prevent it; but a monody 
is one actor's cry from the heart. Doubling in any form is inconceivable. 


(a) The concerted vocal music of the Gabrieli school, with its strong 
element of showmanship, is particularly well suited by a thorough 

Ex. 28. Caccini, Euridice: 

Ex. 29. Monteverdi, Orfeo: 

Exs. 30-31. Peri, Tunestepiaggie": 

i c \ might be performed at end uf pi f 



display of normal ornamentation, as described in Ch. VIII above. 

(b) The new monody of Caccini, Peri or Monteverdi calls for greater 
reticence. But vocal ornamentations are found written out, mainly at 
important closes; and others may be added. 

(127a) Jacopo Peri, UEuridice, Florence, 1600, Preface: 

\ . . Vettoria [sic] Archilei, who has always made my compositions 
worthy of her singing, adorning them not only with those ornaments 
(gruppi) and those long turns of the voice (lunghi giri di voce), simple and 
double, which from the liveliness of her talent are found again at every 
moment, more to obey the fashion of our times than because she con- 
siders the beauty and strength of our singing to consist of them; but 
also [she adorns my compositions] with those pretty and graceful [things] 
which cannot be written, and writing them, cannot be learned from 

(c) The solo songs of the English Elizabethan and Jacobean lutenist 
composers such as Dowland, Campion and Rosseter, and of their 
Caroline and Commonwealth successors, were ornamented freely in 
performance, and some of the ornamentations more or less improvised 
in them were written down in MS. copies. See Appendix III. 

The imitations of Italian monody produced in fair numbers, though 
mostly by minor composers, towards the middle of the seventeenth 
century and subsequently in England can be treated like their models 
in regard to ornamentation. 


(a) The written-out instrumental parts in Monteverdi and other 
early baroque composers invite ornamentation of the same kind as the 
vocal parts, but only where they can indulge in it without getting in the 
way of the vocal parts. Where there are no vocal parts, the character 



of the music must be the guiding factor. A symphonia or ritornello in 
Monteverdi is usually in far too direct and simple a mood to admit 
elaborate or indeed any ornamentation, except that the closes can 
sometimes carry one of the conventional ornamentations approximating 
to what later became the regular baroque trill 

(b) The problem is altogether different in the case of instrumental 
parts not written out but needing to be supplied. 

There are very numerous renaissance descriptions and pictorial 
representations of instrumentalists in combinations sometimes including 
voices and sometimes not. In the former they may be doubling the 
voices, with or without ornamentation; in the latter they may be 
borrowing and probably ornamenting a vocal polyphonic composition, 
they may be playing a composed canzona or ricercare, or they may be 
elaborating their own version of a popular song or dance. The early 
baroque musicians continued all these practices; but they had an added 
problem in providing instrumental parts for the new monodies, written 
out only with their thinly-figured continuo basses (Le. without any 
polyphonic inner parts on which to elaborate), even the choice of 
instruments being, as usual, left to the performers. 

(128) [? Alessandro Guidotti] preface to Emilio de' Cavalieri's Rap- 
presentatione di Anima, e di Corpo (Rome, 1600): 

*A Lyra da Gamba (Lira doppia), a Harpsichord, an. Ar dilute (Chita- 
rone), or Theorbo as it is called, make an excellent effect together; as also 
a smooth Organ [i.e. a small "organo da legno", not a regal with its 
snarling reeds] with an Archlute . . . 

'Signor Emilio [Cavalieri: the composer] would recommend changing 
the instruments about to suit the feelings of the singer. 

'[Awertimentir] The [written-out] Symphonies and Ritornelli may be 
performed with a great many instruments. 

*A violin may double the upper part note for note with excellent 
effect . . .' 

Forty different instruments are listed or mentioned by Monteverdi 
for his Orfeo; but only exceptionally are their names set against an 
actual passage, and only in a few brief but extraordinarily suggestive 
passages are they given written-out parts other than for the straight- 
forward symphonias and ritornelli. These few passages are remarkable 
both for their beauty and their virtuosity, and they afford us a sudden 
glimpse into the tradition of virtuoso improvisation which lay behind 
them. The score of Monteverdi's Combattimento di Tancredi e Clorinda 
is another guide; here the virtuosity is less in evidence, but the writing 
for strings is highly idiomatic. 

Where nothing was provided in writing, the instrumentalists partly 
prepared during rehearsal and partly improvised in performance the 



necessary material which included fully independent as 

as chords to fill up the harmony. The only authorities, apparently, 
tell us in any detail how (and throw light on the previous period, since 
there is clearly no novelty in what they are describing) are Agazzari 
and Praetorius, of whom the latter merely translates the former's 
instructions with some concurring comments of his own. But there are 
many small indications that the method continued, surprisingly, 
throughout the seventeenth century. Doni was complaining In about 
1635 of its difficulties (see App. IV). 
(129) A. Agazzari, Del sonar e sopra "I basso, Siena, 1607, pp. 3, 6, 8, 9 : 

'As foundation, we have the instruments which hold together and 
support the whole consort of voices and instruments in the aforesaid 
Consort, as Organ, Harpsichord etc., and, likewise, in the case of few or 
single Voices, Lute, Theorbo, Harp etc. 

'As ornamentation, we have those which make playful melodies 
(scherzando) and counterpoints (contraponteggiando), thereby rendering 
the harmony more pleasant and resounding; as the Lute, Theorbo, Harp, 
Lirone [bass lyra or lyra da gamba] Cither, Spinet, Chitarrina [small 
Chitarone], Violin, Pandora, and such-like . . . 

'When playing an instrument which functions as foundation, you 
must perform with great discretion, keeping your attention on the body 
of the voices; for if these are numerous, you must play full chords and 
draw more stops, but if they are few, you must reduce the stops and use 
few notes [i.e. in each chord], sounding the work as purely and correctly 
as possible, without indulging much in passage-work or florid movement 
(non passeggiando, 6 rompendo molto); but on the contrary, sustaining 
the voices by sometimes reduplicating the bass in the double-bass register, 
and keeping out of the upper registers when the voices, especially the 
sopranos and the Falsettos, are employing them; at such times you must 
so far as possible try to keep off the same note which the Soprano sings, 
and avoid making florid divisions on it, so as not to reduplicate the voice 
part and cover the quality of that voice or the florid divisions which the 
good singer is making up there; whence it is desirable to play compact 
chords in a lower register (assai stretto, e grave) . . . 

'[Of the instruments of ornamentation:] The instruments which blend 
in with the voices in a variety of ways, blend in with them for no other 
purpose, I think, than to add spice to (condire) the aforesaid consort: a 
harder matter, because as the foundation [instrumentalist] has to play 
[with mainly chordal harmony] the Bass as it stands in front of him, he 
does not need a great skill in counterpoint, whereas [the ornamenting 
instrumentalist] does need it, for he has to compose original parts, and 
original and diverse divisions and counterpoints upon that given bass. 

"Thus whoever performs on the Lute, the noblest instrument of any, 
must perform nobly, with great fertility and diversity, yet not, like 
some who have an agile hand, making continual runs and divisions from 



start to finish, particularly in combination with other instruments all 
doing the same, which leaves nothing to be heard but noise and confusion, 
distasteful and distressing to the hearer. 

'Chords should be struck at times, with restrained reiterations; at other 
times, florid passages both slow and fast should be executed, besides 
[thematic or canonic) points of imitation at different intervals of pitch and 
of time (in diverse corde, e lochi), together with ornaments such asgmppi, 
trilli and accenti* 

'And what we say of the Lute, as chief of instruments, we want applied 
to the remainder each according to its kind, since it would be too long a 
matter to take each individually. 

'Yet since each instrument has its own limitations, the player must 
appreciate them and adapt himself accordingly, in order to succeed. 
Thus: bowed instruments have a different idiom from those sounded by 
a quill or by a finger. Thus he who plays on the Lirone [lira da gamba] 
must take long bows, clear and resonant, bringing out the middle notes 
[in his chords] well, and paying particular attention to the [places requiring 
respectively] Thirds and Sixths: an exacting business, and one which is 
important on this instrument. 

'The violin requires fine passages, long and clear, lively sections 
(scherzi), little replies, and fugal imitations repeated in several places, 
affecting appoggiaturas (accenti), nuances of bowing (? arcate mute), 
gruppi, trilli, etc. 

'The [great bass-biol, the Italian] Violone, proceeds gravely, as befits its 
position in the lowest register, supporting the harmony of the other parts 
with its mellow sonority, keeping on its thick strings as far as possible, and 
frequently doubling the Bass at the Octave below. 

'[Among foundation instruments which can also serve for ornamenta- 
tion are the Theorbo, the Double Harp "as useful in the Soprano register 
as in the Bass" and the Cither and Bass Cither.] But every instrument 
must be used with discretion; since if the instruments are by themselves 
in the Consort, they must take everything on themselves and be the 
leaders of the Consort; if they are in combination, they must have con- 
sideration for one another, and not get in each other's way; if they are 
numerous, they must wait for their own good time and not be like the 
sparrows, all playing at once, and each trying to make the greatest noise.' 

See also the Postscript on p. 116; and App. IV. The following direct 
confirmation of Agazzari may be noted here : 
(129a) Scipione Cerreto, DelVarbore musicale, Naples, 1608, p. 41: 

'The good and perfect player of the cornett must have a good know- 
ledge of the art of counterpoint, so that he can make up varied 
[ornamental] passages at his ease, still more extemporaneously. ... It is 
equally necessary for the trumpet player to know how to make up 






In its melodic aspect, baroque ornamentation continued to develop in 
direct descent from renaissance ornamentation. We merely find an 
extension of the same little runs, changing-notes, turns and other 
figures of the kind. But with the emergence of genuine long appoggi- 
aturas, a harmonic aspect arose, of which the most obvious consequence 
is that the ornamentation cannot any longer be accompanied by the 
plain original, on account of the harmonic disagreement which would 

It is true that Christopher Simpson (Division-Violist, London, 1659) 
still gives instructions in extemporising variations on a reiterated 
ground-bass, with specimens of outstanding virtuosity and poetic 
beauty; while Friedrich Erhard Niedt (Handleitung zur Variation, 
Hamburg, 1706) applies similar methods to a thorough-bass on the 
keyboard. The normal variation form, on the other hand, does not 
reiterate the theme with the variations, even where the same harmonic 
progressions recur. And in the closely related art of more or less im- 
provised ornamentation, the plain original was no longer now in any 
ordinary circumstances intended to support its own ornamental 
elaboration. It was intended to be replaced by it. 


(a) There is no true monody in late baroque music, with the possible 
exception of some French recitative based consciously on verbal 
prosody; but there is a considerable amount of declamatory song or 
arioso which is not quite recitative and not quite aria, and is the nearest 
equivalent to monody then common. PurcelTs famous Expostulation 
to the Blessed Virgin is a splendid example. There is a great deal of 
florid ornamentation in this piece, but it is written out by the composer, 
and there is little if any scope for additional ornamentation from the 
performer. There are also arioso sections in the same piece, but here 
the simplicity of the style precludes florid ornamentation: their point 
is their contrast with the ornate declamatory sections. 

In other examples of the species, the declamatory ornamentation is 



not always as well worked out in writing, so that there are opportunities 
for the performer to elaborate it in the same style on his own initiative, 
(b) In simple strophic songs, or songs with straightforward repeats, 
the principle is to build up interest by gradually increased ornamenta- 
tion. So too with songs in rondo form, and da capo arias. 

(130) B. de Bacilly, VArt de Bien Chanter, Paris, 1668, p. 224: 

'Everyone agrees that the less one can make passages in the first verse, 
the better, because assuredly they prevent the melody being heard in its 
pure form [thereafter in moderation].' 

(131) Francesco Tosi, Opinioni, Bologna, 1723, transl. Galliard, as 
Observations on the Florid Song, London, 1742, p. 93: 

'The manner in which, all Airs divided into three Parts [da capo arias] 
are to be sung. In the first they require nothing but the simplest Orna- 
ments, of a good Taste and few, that the Composition may remain 
simple, plain and pure; in the second they expect, that to this Purity some 
artful Graces be added, by which the Judicious may hear, that the Ability 
of the Singer is greater; and in repeating the Air [i.e. in the "da capo"], he 
that does not vary it for the better, is no great Master/ 

(132) Dr. Burney, Hist, of Music, London, 1776, ed. of 1935, II, 
p. 545: 

'This triffling and monotonous rondo, in which the motivo, or single 
passage upon which it is built, is repeated so often, that nothing can 
prevent the hearer of taste and knowledge from fatigue and languor 
during the performance, but such new and ingenious embellishments as, 
in Italy, every singer of abilities would be expected to produce each 
night it was performed * . / 

Of the two main functions of embellishment, sustaining the musical 
interest and offering scope for the performer's virtuosity, the first is 
well suggested by the quotation above, and the second by the quotation 

(133) Dr. Burney, Hist, of Music, London, 1776, ed. 1935, II, p. 545: 

'Of [Farinelli's] execution the musical reader will be enabled to judge 
by a view of the most difficult [written] divisions of his bravura songs. Of 
his taste [in the technical sense] and embellishments we shall now be able 
to form but an imperfect idea, even if they had been preserved in writing, 
as mere notes would only show his invention, and science, without 
enabling us to discover that expression and neatness which rendered his 
execution so perfect and surprising. Of his shake, great use seems to have 
been made in the melodies and divisions assigned [ascribed] to him; and 
his taste and fancy in varying passages were thought by his contemporaries 



Ex. 32. Farineili: bravura ornamentation of the opening of Quell* 
usignolo from Giacanelii's Nerope, reprinted in Franz Habock, Die 
Gesangskunst der Kastraten, Universal Ed., Vienna, 1923, p. 140 
(this, it may be added, Is a very mild, brief and unelaborate cadenza 
on Farinelli's standards there are much more formidable ones later 
in this song): 

(aj as written 

QucH'u - si-#no- io cue m- na in - 
fl- v as snn% by Farineili <fr i 

QuelTu- si-gno- Io chc in- na mo - r. 

[cadenza as inserted bv Farin 





ir ir <& ir 

[as written] 


(a) Slow movements (often referred to as *the adagio' whether 
actually so marked or otherwise) require ornamentation mainly for 
enrichment and diversification; quick movements ('the allegro*)? if at 
all, mainly for additional virtuosity. 

(134) Joachim Quantz, Essay, Berlin, 1752, partial Engl. transl. as 
Instructions . . . how to introduce Extempore Embellishments, London, 
[c. 1790], p. 3: 

It is a principal Rule with regard to Variations [in the sense of more or 
less improvised ornamentation], that they must have a just reference to 
the plain Air . . . 

'Brisk and lively Variations must not be introduced in an Air that is 
soft, tender and mournful . . . 

'[Variations are] for no other end, than to render an Air in the Cantabile 
Stile more melodious, and [brilliant] Divisions in general more brilliant. 

'[Variations] are only to be introduced after the simple Air has been 
heard first. 

'[For a cumulative effect in sequences:] in the Repetition of one and 
the same [sequential] Passage the Variations ought to be different; which 
is to be observ'd [apart form the special case of sequences] as a general 
Rule [for the avoidance of anticlimax and the gradual building-up of the 



[p. 19:] 'All those instructions concerning Variations are chiefly calcu- 
lated for the Adagio, the most proper and convenient Stile for that 
purpose, where in general most room is left for Embellishments; yet 
there are many among the Examples above, which with equal Propriety 
may also be introduced in the Allegro, the proper choice, however, is 
entirely left to the Discretion of the Performer. 

[Fr. ed., p. 119:] 'The plain melody ought in Allegro as in Adagio to be 
ornamented and made more pleasing by appoggiaturas, and other small 
essential ornaments, according to the demands of the emotion found 

'The Allegro does not admit many arbitrary variations [in the sense of 
more or less improvised ornamentation, as opposed to specific orna- 
ments], because it is most often composed of a melody and passages 
where there is no room for embellishments. But if it is desired to make 
some variations, that must only be in the repeat . . . when by the inadver- 
tence of the composer there are too frequent repeats which might easily 
cause tedium; then the player is obliged to correct them by his talent. I 
say correct; but not disfigure.' 

Ornamentation in quick movements was not, however, confined to 
repeats, though often best so treated. Ex. 33 shows plain and performing 
versions from a Corelli slow movement (grave). Ex. 34 shows the 
opening of a Corelli quick movement (gavotte). 

Ex. 33. Corelli: Op. V, No. 1, (a) in first ed., (b) ornamentation 
from Roger's ed. of 71715: 

,> First 





Ed. of (?) 1715 showing- ornaments as played by Corelli 

Ex. 34. Corelli: versions collated by Sol Babitz and privately 

(a) Plain -version 


f f~ 

. r ij 

r j f i, r , r ., <r ,_q 

(&) Corellis playing version 

- :'.; ' , '... If E J :::,: ,i'J " - 
^ /Tfc f=p. ^ x^ 

tP " * 
(c) C 


s playing- version 

n J J ! r f i ' J ^ r j =i 

0*) I 



playing- version 

.r > iJT 

IT 1 f y+F fr J j 1 J ^ (J i =1 



f-| i ^~j 

J Lr J LflF.-. ..Lr f i|i B r .[. J I 



(135) C. P. E. Bach, Essay, Berlin, 1753, III, 31 : 

"Not everything should be varied, since if it is the repeat will a 

new composition. There are many instances, particularly expressive or 
declamatory passages, which cannot well be varied. Moreover, music in 
the galant style is written with so many new turns and phrases that it is 
not often possible even to understand it at once. All variations must 
accord with the expressive intention of the music, and be as good as the 
original, if not better.' 

(136) Dr. Charles Burney, Hist, of Music, London, 1776, ed. of 
1935, II, p. 10: 

It was formerly more easy to compose than to play an adagio, which 
generally consisted of a few notes that were left to the taste and abilities 
of the performer; but as the composer seldom found his ideas fulfilled 
by the player, adagios are now made more chantant and the performer is 
less put to the torture for embellishments [but see (121) in Ch. VII, 3 

The deeper reason was that galant music grew more chantant with 
the changing artistic climate of the second half of the eighteenth 

(b) Of the following two quotations, the first shows the melodic 
treatment which was still the foundation of later baroque ornamenta- 
tion. If the plain notes are sounded in their original places as here 
described, or if the only liberty taken is to postpone them by some other 
note from the same harmony, bringing in the postponed plain note 
immediately afterwards, it is obvious that the plain and ornamental 
versions will not conflict harmonically, and could in theory still be 
performed simultaneously, although in practice (with such possible 
exceptions as soloists in the tutti passages of concerti grossi) they were 

The second quotation shows what happens when on this melodic 
foundation a harmonic modification is also introduced: chiefly by the 
use of long appoggiaturas. 

(137) Joachim Quantz, Essay 9 Berlin, 1752, Eng. transl. c. 1790, p. 3 : 

'The original Notes may be heard at proper Intervals of Time, as for 
Instance: in varying Notes equal to the Value of a Crochet, the first Note 
of the Variation must for the most Part be the same with the original or 
plain Note, and thus one proceeds with all other Notes of equal Propor- 
tion; or any other Note may be chosen instead of it from the Harmony 
of the Bass, provided the Principal or plain Note be heard immediately 
after it/ 



(138) C. P. E. Bach, Essay, II, Berlin, 1762, XXXII, 5: 

*An intelligent soloist will [make modifications to the harmony] when 
he knows he has an able accompanist ... in introducing embellishments 
. . . The accompanist must modify his harmony accordingly. In 
addition to such substitutions, the accompanist must be attentive and give 
way when embellishments introduced into the solo part lead to a later 
entrance of chords than is actually shown by the figuring/ 



(139) Joachim Quantz, Essay, Berlin, 1752, XVI, 24: 

*In a trio [sonata], little ornamentation must be used, and the second 
part must not be deprived [by excessive ornamentation in the first part] 
of the chance of displaying himself too on his side. The ornamentation 
ought to be such, that it not only suits the subject, but can also be imitated 
by the second part. It should only be used in passages which consist in 
imitations, whether at the fifth above, or at the fourth below, or at the 
same pitch. If the two parts have the same melody one against the other, 
in sixth or in thirds; nothing should be added, unless it has been agreed 
beforehand to make the same variations. 

[25 :] 'If in a trio one of the parts starts to make a variation, the other 
must make it in the same manner; and where the latter is in a position to 
add something further which is suitable and pleasing, it can do so. How- 
ever, it must do it at the end of the variation, so as to show that it is 
capable of imitating simply, as well as of varying further. It is easier to 
make a variation in the first place than to imitate it in the second. 

[26 :] '[In a quartet] there is still less liberty than in a trio to add arbitrary 

[27 :] 'There is more freedom in a concerto with regard to ornamenta- 
tion than in a trio, especially in the Adagio; nevertheless attention must 
be paid to whether the accompanying parts make tuneful progressions, or 
whether they are only the plain harmony. In the first case, the melody [of 
the soloist] must be played quite plainly; but in the second all kinds of 
ornamentation can be made; provided the rules of harmony, taste and 
reason are not violated.' 


While we have every reason to believe that all the baroque national 
schools included a tradition of free embellishment, they did not 
cultivate it in the same degree or altogether in the same form. 

(140) Joachim Quantz, Essay, Berlin, 1752, X, 13: 

'[For learning to play "with a sustained melody"] the French pieces, or 
those composed in this style, are more advantageous than the Italian. For 



the pieces in the French style are mostly pieces characterized composed 
with appoggiaturas and trills, in such a way that hardly anything can be 
added to what the Composer has written; whereas in Music composed 
in the Italian style, a great deal is left to the wishes and capacity of the 
player. In this respect the performance of French music, as written in its 
simple melody complete with its ornaments except for the free embel- 
lishments, is more servile and more difficult than that of the Italian Music, 
as it is written today. However, because for the performance of French 
Music there is no need to know Thorough Bass, nor to understand the 
art of Composition; whereas on the contrary these skills are essential for 
Italian Music; and this because of certain passages which are written in a 
very plain and dry way, to leave the performer the freedom to vary them 
more than once, according to his capacity and judgement, so as continu- 
ally to surprise the hearers by new inventions; for these reasons the Pupil 
ought to be advised, not on any account to undertake to play solos in the 
Italian style before it is time, and before having acquired certain skills in 


(141) EngL translator's/./!., London, 1742, p. 159 to Tosi's Opinion!, 
Bologna, 1723, IX, sect. 48: 

'Many Graces may be good and proper for a Violin, that would be 
very improper for a Hautboy [etc. ; likewise it is wrong] for the Voice, 
(which should serve as a Standard to be imitated by Instruments,) to copy 
all the Tricks practised on the several Instruments. 

(142) P. F. Tosi, Opinioni, Bologna, 1723, EngL transl. London, 
1742, p. 93: 

'The finest Graces [are] confm'd to the most exact Time of the Move- 
ment of the Bass. 

'[13:] Seek for what is easy and natural.' 

(143) Joachim Quantz, Essay, Berlin, 1752, XI, 6: 

'[It is mere bad taste] to load the Adagio with a quantity of ornamenta- 
tion, and to disguise it in such a way that often among ten notes there is 
scarcely one which is in harmony with the fundamental part, and the 
principal melody of the piece can hardly be heard any longer.' 


The following passages were written fairly early in the seventeenth 
century, but have a general bearing on baroque practice as a whole. 
The distinction between those performers with a bent for ornamenta- 
tion, and those without much, if any, bent for ornamentation, is 
suggestive. The basic assumption was that ornamentation was ordinarily 



needed. If the performer was willing and able to produce his own, so 
much the better; if not, someone would have to provide it for him 
even if it had to be the composer himself! Nowadays, it will normally 
have to be the editor. But this emphasises once more how necessary it 
is that a modern editor should provide sample ornamentation where 
needed, except in the somewhat rare instances in which the composer 
has already performed that service for us. 

(144) Bartolomeo Barbarino, // secondo libra delli motetti . . . da 
cant ar si a una voce sola . . . , Venice, 1614, Preface: 

'Because I have heard from many people that in my first book of 
motets . . . some of them are difficult to sing for those who do not have 
the bent to add ornamental passages (passaggiare). Therefore, in this 
second [book] I wanted to write the vocal part in two ways, simple and 
ornamented (passaggiata). The simple for those who do not have the 
bent, and for those who have [a knowledge of] counterpoint and the 
bent, who will be able by themselves to make up ornamental passages 
and the other refinements required for the good manner of singing. The 
ornamented, then, for those who, having the bent, do not have counter- 
point to be able to make up diminutions, as properly one must.' 

(144a) Enrico Radesca, Libra delle canzonette, madrigali et arie, a 
tre, a una, et a due voci, Venice, 1617, Preface: 

* [These compositions] are not given ornamental passages, in order that 
those who by nature are not endowed with the bent may not at all be 
deprived of the work; all the more in that it is clearly seen that, however 
skilful the singer may be, he will never extemporaneously perform that 
ornamental passage exactly as it is written down. ... So those who do 
not have the bent may sing them as they are notated, for which I am 
sure they will not be ungrateful; and those who do have it [the bent] 
will be able to add the ornamental passages to their taste, according to 
their views/ 

(144b) Giovanni Battista Doni, TTrattato della musica scenica' [ca. 
1635], in his Trattati di musica, ed. A. F. Gori, Florence, 1763, II, 69: 

'Nevertheless, if they [ornaments of all kinds] can be allowed in any 
place, this would seem truly to be (contrary to common opinion) in 
theatres, where all sorts of people come together, and the ignorant always 
in greater number than the intelligent; they [ornaments] are better 
adapted to theatrical music than to any other land. ... In chambers, 
similarly, where somewhat delicate music is accustomed to be sung, and 
in gatherings of people who understand music, they [ornaments] are not 
required to be used abundantly, but more sparingly/ 





Free ornamentation began to lose Its status as a primary element of 
interpretation during the second half of the eighteenth century. Its 
decline was one symptom of a profoundly changing attitude, of which 
another symptom was the decline of more or less improvised accom- 
paniment on a figured bass. There can be no doubt of the growing 
desire of that age to have the figuration of the composition decided 
beforehand by the composer. 

Mozart, Beethoven and even Chopin still varied their ornamental 
figuration (particularly of their slow movements) in performance. 
Liszt's variations on the music of other composers, above ail of Chopin, 
were legendary, but are perhaps across the narrow boundary between 
ornamentation and transcription. The art of public improvisation was 
a very flourishing one throughout the nineteenth century, but closely 
allied though it is to the art of freely ornamenting existing music in the 
ordinary course of performance, it is not the same. As a problem of 
normal interpretation, free ornamentation ceases to be of importance 
to us in music of the classical Viennese school. There is scope for it in 
Mozart's operas, but its absence will not leave the music incomplete as 
baroque opera is left incomplete without it. There is scope for it in 
almost all Italian opera in the grand tradition, and here an enterprising 
modern singer might well make some successful experiments. But from 
Haydn and Mozart onwards, if no attempt at all is made to vary the 
notated text beyond the indispensable specific ornaments (chiefly 
appoggiaturas and cadential trills), our interpretation will in no way 
be vitiated. The case is far otherwise with earlier music. 

(145) C. P. E. Bach, Sonatas with Varied Repeats, Berlin, 1760: 

'Nowadays varied repeats are essential, being taken for granted from 
every performer. . . . Often it is nothing but the variations, particularly 
when joined to prolonged and over-curiously ornamented cadenzas, 
which arouse the loudest applause from the audience. And what abuses of 
these two resources occur! . . . Yet in spite of the difficulties and abuses, 
there is always a value in good variation/ 

(146) John Hoyle, Complete Dictionary of Music, London, 1770, ed. 
of 1791 : 



[s. v. 'Adagio':] It is a prevailing custom amongst many performers, 
when they come to an Adagio (as it is slow and consequently easy), to 
throw out favourite passages, which entirely destroy the true harmony 
and intention of the composer. 

'COLORATURA, denotes all manner [i.e. not necessarily improvised] 
of Variations, Trillos, Diminutions, etc., serving to make the Music 

[s.v. 'Expression':] 'What is commonly called good taste in playing or 
singing, has been thought, for some years past, to destroy the true Melody, 
and the intention of the composer. It is supposed by many, that a real 
good taste cannot be acquired by any rules of art; it being a peculiar gift 
of nature, indulged only to those who have naturally a good ear; and as 
most flatter themselves to have this perfection, hence it happens that he, 
who sings or plays, thinks of nothing so much as to make continually 
some favourite passages; but in expressing with strength and delicacy the 
intention of the composer, should his thoughts be most taken up with.' 
[This article is repeated s.v. 'Taste.'] 

(147) Thomas Busby, Complete Dictionary of Music, London [1786], 
s.v. 'Flourish': 

'An appellation sometimes given to the decorative notes which a singer 
or instrumental performer adds to a passage [he wants caution but takes 
the practice for granted].' 


(148) A. E. Miiller, Fortepiano-Schule, ed. Czerny, Leipzig, 1825: 

[Advice to pianists:] 'Arbitrary embellishments are those of which the 
performance is left more to the feeling and taste of the performers. . . . 
To bring them in impromptu is only allowed to those who understand 
composition and have already been trained to this art.' 

(149) Louis Spohr, Violin School, Vienna, 1832, Ch. XXIV: 

[Advice to string quartets:] 'In passages, decidedly solo, the usual embel- 
lishments may be allowed, [cf. Scholes' statement, Oxford Companion to 
Music, s.v. "Fioritura", that the violinist David elaborated the recapitula- 
tions of Haydn's string quartets when leading them.] 

[Ch. XXV] 'Further rules for Orchestra playing are: to avoid every 
addition of turns, doubleturns, shakes, etc. ... in short, every embel- 
lishment properly belonging to the Solo.' 

(150) Nicola Vaccai, Metodo Pratico per il Canto Italiano, Florence, 
c. 1840, Lesson IX: 

[Advice to singers:] 'All the ornaments or so-called embellishments 
which singers are so apt to add to the original melody and accent, are out 
of place and bad.' 



(151) Manuel Garcia, Traitt . . . du chant, Paris, 1847, Ch. Ill, p. 
56, of EngLed. of 1911: 

[Advice to singers :] * When there is no accent to give colour to melody, 
recourse is had to embellishments (orfioriture). This is the case with almost 
all Italian music prior to the present [i.e. the nineteenth] century; for 
authors formerly, in sketching out their ideas, reckoned on the talent of a 
singer to add at pleasure, accent and ornamental accessories. [But many 
of Garcia's musical examples of additional embellishment are applied to 
nineteenth century composers: Rossini, Bellini, etc.] There are different 
kinds of pieces, too, which, from their very nature, must be entrusted to 
the free and skilful inspiration of their executants as, for instance, 
variations, rondos, polaccas, etc. 9 

(152) Henry F. Chorley, Thirty Years' Musical Recollections, 
London, 1862, ed. E. Newman, New York and London, 1926, p. 195: 

[Of Jenny Lind:] 'Her shake (a grace ridiculously despised of late) was 
true and brilliant her taste in ornament was altogether original/ 

Chorley was music critic of The Athaeneum for some forty years till 
his death in 1872. The concert reviews, notices of singers and so on in 
this, the Quarterly Magazine of Music and other nineteenth century 
periodicals often refer to the improvisatory embellishments used by 
singers, and to a lesser extent by instrumentalists. 

(153) The Athaeneum, London, obituary of the violinist, W. Ernst, 
d. 1865: 

'He was not above the virtuoso habit of introducing his own embel- 
lishments into the works of the great masters.' 

(154) F. de Courcy, Art of Singing, London, c. 1868, p. 93: 

[Advice to singers:] 'Embellishments ... are used to give a finish to a 
song; if not found in the music, they should be introduced very cautiously 
and sparingly. ... In Bravura airs alone should the voice be allowed 
freely to use embellishments. [But throughout the book the practice is 
taken entirely for granted.] . . . They depend upon tone, accent, and 
velocity, and their being executed rather sotto voce, and with exquisite 
precision; for though highly elaborated embellishments may surprise the 
ear, yet they never affect the heart. 

[p. 95:] 'In encores . . . change the embellishments, for an artist must 
be barren of invention, indeed, or possess a very poor memory not to be 
ready with a second equally excellent set of graces, etc. 9 

(154a) Stewart Deas, Music and Letters, London, Jan. 1953, p. 8: 

'[On Corelli in apparent unawareness of the incompleteness of his 
slow movements if left unembellished; though the article as a whole is 



very good :] Oddly unconvincing, on the other hand, are most of the 

Sarabands. . . / 

(154b) Dyneley Hussey, Musical Times, London, May 1958, p. 259: 

4 A great singer [Zinka Milanov] even if not always note-accurate so 

that some of her phrases do not correspond with the text.' 



The Cadenza 



A cadenza is an extension of the embellishment outside the time of the 
movement. It occurs at a point where the remaining parts can reason- 
ably wait (except in the case of accompanied cadenzas, which are 
written out, and are not in the strictest sense cadenzas at all). Since it 
usually occurs at or towards the end of a movement or section, it can 
be regarded as a special case of the familiar principle of saving up the 
most striking embellishment for the concluding passage. 

We find incipient cadenzas written out on or around the dominant 
penultimate note of closes in Caccini, Monteverdi and others of their 
school; and similar passages were freely improvised. This remained a 
favourite position, both for voices and for instruments. 

(155) P. F. Tosi, Opinioni, 1723, tr. Galliard as Observations on the 
Florid Song, 1742, p. 128 (with irony): 

'Every Air [da capo aria] has (at least) three Cadences, that are all three 
final. Generally speaking, the Study of the Singers of the present Times 
consists in terminating the Cadence of the first Part with an overflowing of 
Passages and Divisions at Pleasure, and the Orchestre waits; in that of the 
second the Dose is encreased, and the Orchestre grows tired; but on the 
last Cadence, the Throat is set a going, like a Weather-cock in a Whirlwind, 
and the Orchestre yawns.' 

(156) F. W. Marpurg, Beytrdge, Berlin, 1754, p. 15: 

'. . . but before she dies, she gathers all her strength in a cadenza, by 
means of which her heart recovers from its faintness, thus enabling her to 
repeat to her lover, in the Da Capo recapitulation, the first forebodings of 
her just fears/ 

(157) Joachim Quantz, Essay, Berlin, 1752, partial Engl. transl. as 
Instructions, etc., c. 1790, p. 22: 

'The Cadences [cadenzas] to be treated of in this place are properly 
those Embellishments commonly introduced at the End of a Solo Part or 
Air on the last Note but one, mostly on the Fifth of the Key, and are the 
productions of the momentary Invention of the Performer.' 


(158) C. P. E. Bach, Essay, Berlin, 1753, II, Ix, 2ff.: 

* [Of Pauses (fermate):] Their sign is a slur with a dot under it [^], 
which means that a note is to be held as long as required by the character 
of the music. [3 :] A note without the sign may sometimes be held for 
reasons of expression . . . [Pauses are used] on the penultimate note, the 
last note, or the rest after the last [bass] note. In correct usage the sign 
ought to be marked at the beginning and again at the end of an ornamen- 
ted pause [cadenza]. [4:] Pauses over rests occur most often in allegro 
movements and are not ornamented. The other two kinds are usually 
found in slow, expressive movements and must be ornamented if only 
so as to avoid artlessness. In any case, [i.e. even in the absence of a sign] 
elaborate ornamentation is more needed here than in other portions of 



In the Turin MSS. of Vivaldi's violin concertos we find some lengthy 
cadenzas written out, which are not in the printed editions (either 
because Vivaldi wished to keep them for himself., or because he did not 
trust violinists other than his own pupils to meet their formidable 
difficulties including exceedingly high passage-work). 

These violin cadenzas are normally not at the close, but in or before 
the last recapitulation of the thematic material. Coming in such a 
position they contribute strikingly not only to the brilliance but to the 
structure of the movement, especially in their occasional use of thematic 

For wind soloists Vivaldi wrote simpler cadenzas, within the tradi- 
tional stipulation that they shall be able to be taken in a single breath 
for which see (159) and (160) in 4 below. 

This position within the final recapitulation became standard for the 
classical period, when the structural function of the cadenza became 
still more developed. 


Exceptional examples occur (e.g. Bach's third Brandenburg concerto, 
Handel's organ concertos) where with the least hint from the written 
music (in the third Brandenburg, a pair of chords forming a half-close) 
the leader or soloist was required to improvise an entire movement, or 
the solo portions of one. This sets a problem for editor, conductor or 
performer which is usually evaded; but the effect, for example, of 
playing the two Brandenburg chords rnolto espressivo instead of elabor- 
ating them at least to some reasonable duration is quite inadequate. A 
suitable movement must be supplied. 




The following is a brief instrumental example (for a vocal example 
see Ex. 32 in Ch. X, 3, above). 

Ex. 35. J.-B. Loeillet, Sonata F major for flute, violin or oboe, 
Largo, last bars, brief written out cadenza (ed. Hinnenthal, Berlin, 

The performance of such a passage as the above requires all the 
incisive lightness usual in this style of free ornamentation. A baroque 
cadenza, whether vocal or instrumental, should always sound as if 
improvised, whether it is so in fact or not. It is there not to add weight 
to the music, but a last touch of exhilaration. 

(159) Joachim Quantz, Essay, Berlin, 1752, partial Engl. transl. as 
Instructions, etc., c. 1790, p. 25: 

'Regular Time is seldom to be observ'd in Cadences. . . . Those for 
Voice or Wind Instruments ought to be short and so managed that they 
may be perform' d in one Breath, but those for String Instruments are 
not limited, but the Performer has as much Latitude given him, as his own 
Skill and fruitfulness of Invention will permit, but notwithstanding will 
gain more Applause from the Judicious by a moderate length than other- 


[Fr. ed. p. 162:] *If it is easy to invent cadences in two parts and write 
them down; it is nevertheless very hard to make them on the spot without 
having agreed beforehand. However if the advantages offered by imita- 
tion and the practice of dissonance are familiar, these difficulties can be 


(160) Dr. Burney, Music in Germany, London, 1773, II, p. 153: 

'The cadences which his majesty [Frederick of Prussia, Quantz's 
employer] made, were good, but very long and studied. It is easy to 
discover that these concertos were composed at a time when he did not so 
frequently require an opportunity of breathing as at present; for in some 
of the divisions, which were very long and difficult, as well as in the 



closes [cadenzas], he was obliged to take his breath, contrary to rule, before 
the passages were finished. 

[p. 233:] 'if any of his Italian troops dare to deviate from strict discipline, 
by adding, altering or diminishing a single passage in the parts they have 
to perform, an order is sent, depar le Roi, for them to adhere strictly to the 
notes written by the composer, at their peril. This, when compositions 
are good, and a singer is licentious, may be an excellent method; but 
certainly shuts out all taste and refinement/ 

(161) William Jackson, Observations on the Present State of Music 
in London, London, 1791, pp. 20-23: 

'The Performer, no doubt, ought to be able to run from the bottom to 
the top of the keys in semitones; but let him be satisfied with having the 
power, without exerting it. ... Cadences with, for ever, a concluding 
shake [trill] though sometimes it seems as if it would never conclude. . . 

(162) Dr. Burney, Commemoration of Handel, London, 1785: 

*One night, while Handel was [at the harpsichord] in Dublin, Dubourg 
having a solo [violin] part in a song, and a close to make ad libitum, he 
wandered about in different keys a great while, and seemed indeed a 
little bewildered, and uncertain of his original key; but at length, coming 
to the shake [trill] which was to terminate this long close, Handel, to the 
great delight of the audience, and augmentation of the applause, cried 
out, loud enough to be heard in the most remote parts of the theatre, 
"You are welcome home, Mr. Dubourg"/ 





An ornament is a short melodic formula which has formed in the 
tradition of free ornamentation as a crystal forms in a saturated 

Oriental music is full of ornaments. They were common in medieval 
music, where signs for them existed. They are not very conspicuous in 
renaissance music, though the later sixteenth century treatises describe 
a few rather fluctuating ones, not sharply defined. They reached a peak 
in baroque music, when an immense vocabulary of signs developed. 
These signs, the names attached to them and the interpretation allotted 
to them all show an extreme inconsistency, and the reader is warned 
that the same sign or the same name may have many different meanings 
in different sources, indeed often in the same source. Much of this 
inconsistency cancels itself out, however, and it is not impossible to 
trace certain principles fundamental enough to give the musician what 
he chiefly needs : a clear account of the standard practices underlying 
the countless variations of detail. 

These standard practices are adequate to ninety-nine situations out 
of a hundred. In the hundredth situation, the ordinary rules may make 
intolerably bad grammar or excessive harshness, and we have to reach 
the best compromise we can. In certain styles a knowledge of some of 
the more exotic ornaments is necessary. But in most baroque music it 
is the few plain and uncontroversial ornaments, not the many compli- 
cated variants, which are needed. 


(a) There are obligatory and optional ornaments. 

(163) Jean Rousseau, Traite de la mole, Paris, 1687, p. 72: 

'You must practice all the ornaments in all their fullness, especially the 
prepared trill and the appoggiatura/ 

The cadential trill, with its preparation (accented upper-note start), 
and the (long) appoggiatura are the only baroque ornaments of which 
it would be true to say that in normal contexts they are obligatory, as 
is seen in Ch. XX and Ch. XIV below. Hotteterre's list (Principes 



de la Flute ^ 1707, p. 21) of 'ornaments which, are absolutely necessary 
to the perfecting of the performance' also includes these two, among 
others whose necessity is in fact less unconditional 

(164) F. W. Marpurg, (a) Historisch-kritische Bey tr age, Berlin, 
1756, III, 2 and (b) Ankitung, 2nd ed., Berlin, 1765, I, ix, 3, p. 44: 

[a:] 'Regarding the smaller ornaments, they are so essential at most 
places that without their strict observance no composition can please the 
more refined ears. 

[b:] But where does one learn what notes are given ornaments or at 
which point of the melody this or that ornament ought to be introduced? 
One should hear persons who are reputed to play elegantly, and one 
should hear them in pieces one already knows. In this way one may form 
one's taste, and do likewise. For it is impossible to devise rules to meet all 
possible cases, so long as music remains an inexhaustible ocean of options, 
and one man differs from the next in his appreciation. One should try to 
give a piece of music in which ornaments have not yet been marked to 
ten different people, each playing in the good style of the day, and ask 
them for their ornaments. In certain cases, perhaps, many will agree; in 
the rest, they will all be different.' 

(b) The absence of a sign does not preclude an ornament. 

If the context implies an obligatory ornament, the mere absence of 
a sign does not absolve the performer from the necessity of introducing 
it. Indeed, the very fact that the ornament is obligatory made its 
presence so obvious that the sign is more likely to be absent. 

If the context suggests an optional ornament, the mere absence of a 
sign is no reason for not introducing it. A baroque performer did not 
wait to be told when he could put in an ornament. See the previous 
quotation (164b). 

(c) The presence of a sign does not enforce an ornament. 

The corollary of the above is that if a sign is present but the per- 
former does not wish to put in an ornament, he is not compelled to do 
so unless the ornament in question is in any case an obligatory one. 
This also is implied in (164b). 

If, however, he does wish to put in an ornament, and the sign is 
specific, he is still not compelled to put in that particular ornament, 
although he will normally give it preferential consideration. 

It may also happen that the sign is not specific. It is quite probable 
that the single and double strokes used by the English virginalists at 
the turn of the sixteenth and seventeenth centuries were not specific, or 
at any rate not wholly so. The common + and x of the main baroque 
period were certainly not very specific. The + is used in the violin 
part of Mondonville's Op. Ill, of which the harpsichord part carries 



more differentiated signs, the significant instruction 
'ornament like the harpsichord'. The inference is that a violinist, even 
in France, would not have understood the more differentiated signs, 
which were, in fact, mainly developed by harpsichordists. 

The best general approach to the baroque signs for ornaments is to 
treat them not as commands but as hints. 

(d) But certain schools require a stricter treatment of the ornaments. 

(165) Fr. Couperin, Pieces, Liv. Ill, Paris, 1722, preface: 

I am always surprised (after the pains I gave myself to mark the orna- 
ments which suit my Pieces, of which I have given, incidently, a clear 
enough explanation in my special Treatise, known under the tide of The 
Art of playing the Harpsichord) to hear of people who treat them with- 
out respecting them. This is an unpardonable neglect, in view of the fact 
that it is not at all an optional matter to take such ornaments as one 
wishes. I declare then that my pieces must be performed as I have marked 
them, and that they will make a certain impression on people of good 
taste only if everything which I have marked is observed to the letter, 
without addition or subtraction.' 

Yet Couperin himself marked, for example, his first (string) version 
of *Les Pellerines' only with a few crosses, but his second (harpsichord) 
version with all his customary explicitness. No doubt it was on the 
harpsichord that he regarded his exact ornaments as so particularly 
integral to the music. It is worth trying them all, even if a few are sub- 
sequently dropped. The crystalline sparkle with which they invest not 
only the harpsichord music but also the vocal and the chamber music 
of this fascinating school is so characteristic that it well rewards the 
initial effort; and once mastered they come easily and gracefully 

(166) C. P. E. Bach, Essay, Berlin, 1753, II, i, 4: 

'In justice to the French it should be said that they mark their orna- 
ments with scrupulous exactness. So too do the masters of the harpsichord 
in Germany, yet without carrying their ornaments to excess. Who can 
tell but that our restraint both in the quantity and in the quality of our 
ornaments has been the influence which caused the French to leave their 
earlier way of putting on ornament on almost every note, to the detri- 
ment of clarity and of noble directness? 

[25:] 'Since our present taste, to which the Italian lei canto has contri- 
buted so much, requires other ornaments in addition to the French, I have 
had to draw on the ornaments of various countries. I have added some 
new ones to these. I believe that the best style of performance, on what- 
ever instrument, is the one which skilfully unites the accuracy and bril- 
liance of the French ornaments with the smoothness of Italian singing. 
Germans are in an excellent position to accomplish such a union . . .' 



(167) F. W. Marpurg, Bey tr age, Berlin, 1754, p. 27: 

'Many of our most famous players admit that they have learnt the 
nicety of their execution from the French. In the notation of the orna- 
ments with which a piece is to be played, they prove more careful than 
other nations. Among the Germans, next to the famous Georg Muffat 
[c. 1645-1704], it seems to have been the able harpsichordist Joh. Gasp. 
Ferdin. Fischer [c. 1665-1746] who first spread this kind of music among 
us. He has ornamented in the French manner all his keyboard scores, e.g. 
those in the musikalischen Bhimenbikhlein which appeared at the beginning 
of this century, and one sees from the explanations prefacing the pieces 
that these notations must have been fairly unknown then/ 

(e) Ornaments must suit their contexts. 

(168) Jean Rousseau, Traite de la Viole, Paris, 1687, p. 73: 

'Avoid a profusion of ornamental figures, which only confuse the 
melody and obscure its beauty/ 

(169) F. W. Marpurg, Critische Musicus, Berlin, 1750: 

'[The Berlin School] makes restrained use of ornaments and ornamenta- 
tion, but what is used is the more choice and the more beautifully and 
clearly performed/ 

(170) Joachim Quantz, Essay, Berlin, 1752, VIII, 19: 

'It is true that the ornaments of which we have spoken are absolutely 
necessary for good expression. All the same they must be used in modera- 
tion, it never being right to go to excess in what is good/ 

(171) C P. E. Bach, Essay, Berlin, 1753, II, i, 9: 

'Above all, the extravagant use of ornaments is to be avoided. They 
are to be looked upon as spices which can spoil the best dish, or as details 
which may ruin the most admirable building. Notes of no particular 
significance, or sufficiently brilliant in themselves, should be kept free of 
ornaments, for the only function of these is to increase the prominence 
and emphasis of notes and to make them stand out from the others/ 

(172) C. P. E. Bach, Essay, Berlin, 1753, II, i, 8: 

'The expression of simplicity or of sadness admits o fewer ornaments 
than other feelings. [10:] Many passages permit of more than one kind of 
ornament. In such cases, the art of contrast may well be brought into 
play; use first a tender ornament, next a brilliant one, or even . . . play 
the notes simply as written. [19:] All ornaments are in a proportionate 
relationship to the length of the main note, the speed, and the feeling of 
the music ... the more notes there are in an ornament, the longer the 



main note must be ... (avoid] ornaments of too many [or] excessive 
embellishment of rapid notes.' 

(f ) Ornaments are influenced by the in use. 

(173) C P. E. Bach, Essay, Berlin, 1753, III, 7: 

fi The ear accepts more movement from the harpsichord from 

other instruments/ 

This implies, what experience confirms, that more frequent and more 
brilliant ornaments can be successfully used on the harpsichord than 
on most other instruments, including, for example, the piano; and this 
is a factor which should be taken into account. 

(g) In fugal entries, etc., the use of ornaments must be consistent. 

(174) Joachim Quantz, Essay, Berlin, 1752, XIII, 29: 

'In passages which resemble one another, the [ornamental] variations 
ought always to be the same/ 

(175) C. P. E. Bach, Essay, Berlin, 1753, II, i, 28: 

'All imitations must be exact down to the last detail. Hence the left 
hand must practice ornaments until it can imitate them efficiently/ 

If 3. S. Bach, for example, marks one sign over one fugal entry and 
another over another fugal entry in the same composition, this is not 
a distinction subtly intended, but simply casual notation of the common 
baroque kind. The same ornament should be played. 

Even in imitations, it may be enough to reproduce the general 
contour of the ornament first heard: e.g., a short trill in an upper part 
may be echoed by a turn to avoid clumsiness in a bass imitation of it. 

(h) The ornaments on sequences should often be cumulative. 

This is to avoid anticlimax, and is the corollary of the above: see 
(134) in Ch. X, 3 above. 

(i) Ornaments in ensemble music should be concerted. 

(176) J. A. Scheibe, Critische Musicus, 2nd ed., Leipzig, 1745, art. 78 : 

*The conductor must see to it that all the violins use the same orna- 
ments as their leader/ 

(j) Ornaments must be supplied with accidentals at discretion. 

(177) C. R E. Bach, Essay, Berlin, 1753, II, i, 17: 

'The notes of an ornament adapt themselves to the sharps and flats of 
the key signature [with exceptions and modifications] which the trained 
ear at once recognises/ 


(k) Baroque ornaments never anticipate the beat. 

There are ornaments, such as the turn, which fall between beats, but 
no standard ornament attached to a beat which anticipates that beat. 
(178) C. P. K Bach, Essay, Berlin, 1753, II, i, 23: 

'All ornaments written in little notes relate to the following note. Thus 
while the previous note is never curtailed, the following note loses as 
much of its duration as the little notes take away from it ... 

[24 :] 'Following this rule, the little notes rather than the main note are 
sounded with the bass and the other parts. The performer should carry 
them over smoothly into the following note. It is a mistake to arrive 
with an accent on the main note . . . 

[Ed. of 1787 adds:] It might be thought unnecessary to repeat that the 
remaining parts including the bass must be sounded with the first note of 
the ornament. But as often as this rule is invoked, so often is it broken.' 


(179) C. P. E. Bach, Essay, Berlin, 1753, II, i, 1: 

'[Ornaments] join notes; they enliven them . . . they give them 
emphasis and accentuation . . . they bring out their expression.' 

These functions clearly overlap; but there is a distinction implied 
here which is valuable for interpretation. For if we know what an 
ornament is put there to do, we are more than half way to knowing 
what to do with the ornament. 

(a) There are primarily melodic ornaments whose natural function is 
to 'join notes' and 'enliven them'. Turns are an excellent example; so are 
the little scale-wise passages sometimes treated as a specific ornament 
under the name of tirata. Such ornaments tend to be: either on or 
between beats; little accented; long and smooth enough to fill much 
or all of the duration of their main notes. 

(b) There are primarily rhythmic ornaments whose natural function 
is to 'give emphasis and accentuation'. Mordents, slides and acciacca- 
turas are typical examples. Such ornaments are characteristically on 
the beat (or on a syncopation displacing the beat); highly accented; 
crisp and short, or at most prolonged to take no very great proportion 
of their main notes, 

(c) There are primarily harmonic ornaments whose natural function 
is to 'bring out the expression'. Cadential trills and long appoggiaturas 
are the main examples. Such ornaments are necessarily on the beat; 
firmly accented or with a languishing stress, according to the feeling of 
the passage; long enough to delay their resolution on to their main 



notes expressively, and at times to push those 

own beat and into the space afforded by an ensuing rest, 

(d) Where more than one of functions Is in evidence, the 

disposition of each has to be reconciled. A long mordent, for example, 
must start with enough incisiveness to fulfil its rhythmic function, but 
continue with enough grace to fulfil its melodic function; a trill 
must have its initial upper note stressed and prolonged to 

fulfil its harmonic function, but its continuation smooth enough to 
fulfil its melodic function. And so with other combinations of function. 


(See also the Index of Signs on pages 571-578.) 
A: the Appoggiatura Family 

A (i): The Appoggiatura proper 

(a) early baroque (indeterminate length) 

(b) long 

(c) short 

A (ii) : The Compound Appoggiatura (or disjunct double 


A (iii) : The Slide (or conjunct double appoggiatura) 
A (iv) : The Acciaccatura (crushed appoggiatura) 

(a) simultaneous 

(b) passing 

A (v): The Passing Appoggiatura 

B: the Shake Family 

B (i) : The Tremolo (or organ shake) 
B (ii) : The Vibrato (or close shake) 
B (iii) : The Trill (or shake proper) 
B (iv) : The Mordent (or open shake) 

C: the Division Family 

C(i): Passing Notes 
C (ii) : Changing Notes 
C (iii) : Turns 
C(iv): Broken Chords 
C (v) : Broken Notes 
C(vi): Broken Time 


D: Compound Ornaments 

D (I) : Appoggiatura with 

(a) trill 

(b) half-shake 

(c) mordent 

(d) arpeggio 

(e) turn 

D (ii) : Ascending turn (slide with turn) 
D (iii): Trill with 

(a) mordent 

(b) turn 

D (iv) : Ascending trill (slide with trill) 

D (v) : Descending trill (turn with trill) 

D (vi) : Double Cadence 

D (vii) : Double Relish 

D (viii) : Truncated Note with other ornaments 



A (i): The Appoggiatura Proper 

English: Appoggiatura 


Forefall, Backfall, Half-fall 



Latin: Accent us 

Italian: Accento 



French: Accent 



Cheute or Chute 


German: Accent 




The Italian verb appoggiare means 'to lean' and implies an ornamental 
note expressively emphasised and drawn out before being more gently 
resolved on to its ensuing main note. 

There is no appoggiatura so characteristic of this family of ornaments 
as the long appoggiatura, and no ornament more characteristic of its 
age. Its influence on the history of music surpasses that of any other 
ornament, and it is an outstanding instance of the principle: the 
impromptu ornamentation of today becomes the written figuration of 
tomorrow. Long appoggiaturas fully written out in the notated text are 
so indispensable to nineteenth-century melody and still more harmony 
that their origin in more or less impromptu ornamentation is apt to be 
forgotten. The harmonic progressions with which Tristan opens 
perhaps the most revolutionary passage in music since Monteverdi, 
and the most decisive for future developments are not merely con- 
ditioned but created by the bold and original handling of written-out 
long appoggiaturas. Mahler without his written-out long appoggiaturas 
would be deprived indeed. No one who has absorbed these realities 
into his musical consciousness will ever again wish to deprive baroque 



music of its long appoggiaturas merely because they were not written 
down. They were intended, they were familiar to contemporary com- 
posers and performers alike, and they add profoundly to the expressive- 
ness of the passages which demand them. There are very few details of 
authentic interpretation which are so important as the correct and 
imaginative use of long appoggiaturas. 


(a) The name appoggiatura itself does not carry any implication as 
to the direction of movement of the appoggiatura. 

We use the descriptions 'upper' or 'superior' or 'descending appoggi- 
atura' for a stepwise downwards movement from the tone or semitone 
above (Old Eng. backfall', Fr. coule, chute, port-de-voix, etc., en descen- 
dant, coulement\ Ger. Accent f attend). 

And we use the descriptions 'lower' or 'inferior' or 'ascending 
appoggiatura' for a stepwise upwards movement from the tone or 
semitone below (Old Engl forefall\ Fr. cou/e, etc., etc., en montant; 
Ger. Accent steigend). 

(b) Except in connection with recitative, it is most usual now to 
think of appoggiaturas as moving by step: but this was less so in the 
baroque period. 

Ex. 36. Johann Mattheson, Vollkommene Capellmeister, Hamburg, 
1739, II, iii, 24 and 25, appoggiaturas by leap: 


Q ' L . 

(c) Rules are found for the movement of appoggiaturas., but they 
are far from reliable, and the performer has in practice to judge each 
context on its merits. 

(180) Joachim Quantz, Essay, Berlin, 1752, VIII, iii: 

'[An appoggiatura, when] a retardation from a previous note . . . may 
be taken from above or below. . . . When the previous note is one or two 
steps higher than the following note ... this must be taken from above. 
When the previous note is lower, this must be taken from below/ 

(181) C. P. E. Bach, Essay, Berlin, 1753, II, ii, 6: 

'An appoggiatura may be a repetition of the note before or otherwise 
... [it may proceed by step, ascending or descending, or by leap].' 


Ex. 37. C. P. E. Bach, appoggiatura by leap (my 

^ < : 

This last, like most appoggiaturas which move by leap, is a note of 
retardation (in exactly the complementary sense to a note of anticipa- 
tion). But stepwise appoggiaturas show much less predilection for being 
notes of retardation (i.e. prepared). Nor do they necessarily approach 
their main note from the same side as the previous note, though they 
usually do so. 

(182) Florido Tomeoni, Theorie de la Musique Vacate, Paris, 1799, 


(d) The movement of appoggiaturas is, however, subject to the 
ordinary rules of musical grammar. 

(183) F. W. Marpurg, Anleitung, 2nd ed., Berlin, 1765, I, k, 9, p. 49: 

'One must no more introduce faulty progressions with an appoggiatura 
than one may count on an appoggiatura to save consecutive fifths/ 


(184) Francois Couperin, UArt de toucher le Clavecin, Paris, 1716, 
ed. of 1717, p. 22: 

'Strike [appoggiaturas] with the harmony, that is to say in the time 
which would [otherwise] be given to the ensuing [main] note/ 

(185) F. W. Marpurg, Anteitung, 2nd ed., Berlin, 1765, I, ix, 4, p. 48: 

'All appoggiaturas . . . must come exactly on the beat . .' . 

[8] 'If several parts are present with an appoggiatura, the accompanying 
parts must not be delayed by that, but should be played at once with the 
appoggiatura, only the main note to which the appoggiatura is the 
accessory being delayed.' 


(185a) Joachim Quantz, Essay, Berlin, 1752, VI, i, 8: 

'It is a general rule, that one must make a small [N.B.] separation 
[silence of articulation] between the appoggiatura and the note which 



proceeds It, above all if the two notes are at the same pitch; so that one 
can make the appoggiatura distinctly heard/ 

(186) C. P. E. Bach, Essay, Berlin, 1753, II, ii, 7: 

'All appogglaturas are performed more loudly than the ensuing note, 
including any ornaments which it [the ensuing note] may carry, and are 
joined with it, whether slurs are written or not.* 

See also (195) and (196) in 6 below. 


The appoggiatura occurs, as one of a group of similar little figures 
crystallizing out of the main tradition of free ornamentation and 
together called accenti, early in the baroque period. 

Ex. 39. M. Praetorius, Syntagma, III, Wolfenbiittel, 1619, p. 233: 

f ._! 



The following in Frescobaldi are written out: 

Ex. 40. G. Frescobaldi, 'Partite sopra 1'aria della Romanesca', 

'Fieri musical!* "Avanti il recercar" 

We have here, then, a genuine appoggiatura of moderate length: 
i.e. taking a third of dotted notes and a quarter of undotted notes (no 
doubt with some variability in performance). 

Such an appoggiatura is long enough to be heard as a momentary 
diversion of the harmony, and has the effect of an accented passing 
note. It does not actually change the harmonic progression. It will be 



accented and louder than its ensuing note. It will be on 

the beat. 

Through most of the seventeenth century the evidence 
this was then the standard appoggiatura, 

When ascending it is called beat by Playford (Introd., on) and 
Simpson (Division-Violist, 1659); half-fall fever from a Half-Note 
beneath') by Mace (Mustek" s Monument, 1676); by Locke 

(Melothesia, 1673) and Purcell or Parcel's editor (posthumous Lessons, 
1696): forefall becoming the standard English term. When descending 
it was called backfall. 

Ex. 41. J. Playford, Introduction, London, 1654 (1660 on), appog- 
giaturas : 

a beat 

Ex. 42. Appoggiaturas in H. Purcell's posthumous Lessons, 
London, 1696: 

aforcfaU a baclcfall 

This appoggiatura of moderate length did not disappear from later 
baroque music, and in some contexts it is the obvious solution, as either 
melodic or more probably harmonic considerations will suggest. But it 
ceased to be standard, and was increasingly over-shadowed by two 
other appoggiaturas never moderate or indefinite in length, but in the 
one case decidedly long, and in the other case, decidedly short. 


(a) From the last years of the seventeenth century onwards, we meet 
evidence to suggest that the standard appoggiatura now took half the 
length of an undotted main note; two-thirds of the length of a dotted 
main-note; all the first of two tied notes in compound triple time; and 
all of a note before a rest. 

Ex. 43. D'Anglebert, Pieces de Clavecin^ Paris, 1689, appoggiaturas 
(cheute ou port de voix): 

Cheute on port de 
voix en montant 



Ex. 44. J. M. Hotteterre, Principes de la fltite traversiere, Paris, 


This notation by little notes apparently showing the length required 
is found elsewhere, but never became standard practice, and is all too 
frequently deceptive see (b) above. The vast majority of little notes 
(mostly written as quavers) used to show appoggiaturas in later baroque 
music bear no relation to the intended length, which is frequently 
much greater and sometimes shorter. (The stroked quaver J 1 was not 
a baroque sign in the nineteenth century sense of a short appoggiatura, 
though often used, very incorrectly and misleadingly indeed, in modern 
editions of baroque music.) 

Ex. 45. Dieupart, Suites de Clavecin, Paris, c. 1720, appoggiaturas: 

Port de voix 
Forefal) up 


up Backfall - ; [-. 

Ex. 46. J. S. Bach, Clavier-Buchlein, begun Cothen, 1720: 

(187) Joachim Quanta, Essay, Berlin, 1752, VIII, Iff.: 

'Hold the appoggiatura half the length of the main note. [But] if the 
appoggiatura has to ornament a dotted note, that note is divided into 
three parts, of which the appoggiatura takes two, and the main note one 
only: that is to say, the length of the dot . . . 

'When in six-four or six-eight time two notes are found tied together, 
and the first is dotted ... the appoggiaturas should be held for the length 
of the first note including its dot. . . . When there is an appoggiatura 
before a note, and after it a rest; the appoggiatura ... is given the length 
of the note, and the note the length of the rest/ 


Ex. 47: 

,""'&" ;v '- " * "' TT-H^T <>- M T 

~\5? 1 

frtr _ ^.r:Fz:| 




^^^fe^ 1 ^^ 

g> ft I t -/ !?>T-^- 


^trr^^ * " ' 

-^ .'jn I, ,7.1 1 ..I __ , 


(188) C. P. E. Bach, 

j&^oy, Berlin, 

Mir ?-*-ft 

1753,11, ii, 11: 

'The general rule for the length of the appogglaturas Is to take from the 
following note, if duple, half its length; and if triple, two-thirds of its 

(b) Appoggiaturas of more than standard length also became 
fashionable during the first half of the eighteenth century. 

(189) Fr. Geminiani, Art of Playing on the Violin, London, 1751, 
p. 7: 

'The Superior Apoglatura is supposed to express Love, Affection, 
Pleasure, etc. It should be made pretty long, giving it more than half the 
Length or Time of the Note it belongs to. ... If it be made short, it will 
lose much of the aforesaid qualities; but will always have a pleasing 
Effect, and it may be added to any note you will. 

'The Inferior Apogiatura has the same Qualities with the preceding . . / 

(190) GaUiard, f.n. to Tosi's Opinion!, 1723, tr, as Observations on 
the Florid Song, London, 1742, p. 32: 

'You dwell longer on the Preparation [i.e. the appoggiatura] than on 
the [main] Note for which the Preparation is made/ 

(191) C. P. E. Bach, Essay, Berlin, 1753, II, ii, 16: 

'There are cases in which [the appoggiatura] must be prolonged beyond 
its normal length for the sake of the expressive feeling conveyed. Thus 
it may take more than half the length of the following note. Sometimes 
the length is determined by the harmony.' 

(c) The last sentence in the above quotation is important. 

The likeliest interpretation of an eighteenth century appoggiatura, 
and the one to try first when in doubt, is long. It may alternatively be 
genuinely short. But if it is long, its length must ultimately be decided 
by context rather than by rule. 

Arnold Dolmetsch (Interpretation of the Music of the llth and l&th 
Centuries, London, 1915, 115-116) gives an instructive instance in J. S. 



Bach's Prelude XVIII from Vol. II of the Torty-Eight'. The rule would 
put the first pair of appoggiaturas at a crotchet length, their main notes 
being pushed over into the ensuing rest. That makes nonsense of the 
harmony. The next pair, having no ensuing rest, must be of quaver 
length; later, in bars 44-45, the same 'appoggiaturas* are written out 
as quavers. The solution throughout is therefore a quaver length. 

Ex. 48. J. S. Bach, appoggiaturas: 

(d) A long appoggiatura on an exceptionally long main note may 
need shortening a little; but it is generally desirable to make long 
appoggiaturas as long as the rules and the context make feasible. 

(e) Appoggiaturas taking three-quarters of the length of an ordinary 
duple note were beginning to be regarded as regular by the middle of 
the eighteenth century, though the standard length continued to be 
one half. 

Ex. 49. Leopold Mozart, Violinschule, Augsburg, 1756, IX, 4, 
very long appoggiaturas : 

ll :> J. . 1 



\?'* 1 * 




LJ; Q_J 

(f) Dotted figures in baroque music tend to have their dotted notes 
prolonged (see Ch. XLIII below); this has the effect of prolonging an 
appoggiatura on such notes. 

(192) Joachim Quantz, Essay, Berlin, 1752, partial Eng. tr. c. 1790, 
p. 18: 

'The short note after the Point [dot] is always to be play'd very quick. 
An Apogiatura prefix'd to a pointed Note, must be play'd exactly to the 
time of the larger or principal Note, and the latter to the time of the 
Point, and be play'd softer than the former.' 

C. P. E. Bach prints an interesting example of this. 



Ex. 50. C. P. E. Bach, Essay, Berlin, 1753, II, ii, 11, 
on a dotted note: 

(g) The last words In the last quotation above refer to the stress on 
the appoggiatura, and the diminution of volume Into and upon its 
ensuing main note, which were characteristic of long appoggiaturas. 
This Is in conformity with the normal baroque (and Indeed Intrinsic) 
diminuendo from discord on to resolution, and Is both natural and 
expressive. Some writers describe a crescendo on the appogglatura 
Itself, If long enough; the effect in practice is a mild and somewhat 
gradual sforzando, but the least exaggeration will be fatal here. 

(193) Galliard,/./*. to Tosi's Opinioni, 1723, Eng. tr. as Observations 
on the Florid Song, London, 1742, p. 32: 

'You lean on the first [i.e. the appoggiatura] to arrive at the [main] Note 

(194) Fr. Geminiani, Art of Playing on the Violin, London, 1751, 
p. 7: 

'Swell the Sound [of the long appoggiatura] by Degrees.* 

(195) F. W. Marpurg, Ankitung, 2nd ed., Berlin, 1765, p. 48: 

'The note with which the [appoggiatura] is made should always be 
sounded a little louder than the main or essential note, and should be 
gently slurred towards it/ 

(196) Joachim Quantz, Essay, Berlin, 1752, VIII, ill; 

'Swell [the appoggiaturas] if time permits and slur the ensuing note to 
them somewhat more softly/ 



(197) Joachim Quantz, Essay, Berlin, 1752, VIII, i: 

'Ports de voix, in Italian Appoggiaturas, are not only ornaments, but 
necessary elements. Without them the tune would often be very dry and 
very plain. For a melody to have a stylish air, it must always have more 
consonances than dissonances. However, when there are several con- 
sonances in succession, and after some quick notes there comes a long 
consonance, the ear can easily get tired of it. Dissonances are sometimes 
needed then to stimulate and awaken it. This is where the appoggiaturas 



can make a great contribution; for they turn into dissonances, as fourths 
and sevenths, when they are before the third or sixth counting from the 
principal note [bass of the harmony], but are resolved by the following 

(198) C. P. E. Bach, Essay, Berlin, 1753, II, ii, 1: 

'Appoggiaturas are among the most necessary ornaments. They enrich 
the harmony as well as the melody. They enhance the appeal of the 
melody by joining notes smoothly together, and with notes which 
might be found tedious on account of their length, by making them 
shorter while satisfying the ear with sound. At the same time they make 
other notes longer by repeating, on occasion, the note before. . . . 
Appoggiaturas change chords which in their absence would be too 
straightforward. All syncopations and dissonances can be attributed to 
them. Where would the art of harmony be without these ingredients?' 

(199) F. W. Marpurg, Anleitung zum Clamerspielen, 2nd ed., Berlin, 
1765, 1, ix, 1, p. 46: 

'The appoggiatura (Vorschlag) is also called suspension (Vorhalt), and 
consists, as implied, in suspending a note by a previous one. It is notated 
either by certain signs or by little subsidiary notes, or properly written 
out. The first notation is no longer usual, or can only be used with the 
very shortest appoggiaturas. Formerly, we used for it a simple cross, a 
hook before the note, or a small, slanting line. When, later on, long 
appoggiaturas occurred, we began to introduce the small, subsidiary 
notes, and with them the second notation of this ornament/ 

It is only long appoggiaturas which behave like true suspensions and 
change the progression of the harmony. 


(a) The short appoggiatura varies in length from the shortest perform- 
able, at the minimum, to a quarter or more of its main note according 
to the context, the maximum being, however, always shorter than would 
sound like a long appoggiatura in that context. 

(200) Joachim Quantz, Essay, Berlin, 1752, VIII, ii: 

'Appoggiaturas are marked with little added notes, so as not to con- 
fuse them with the ordinary notes, and they take their value from the 
notes before which they are found. It does not much matter whether they 
have more than one tail, or whether they have none. It is, however, 
usual to give them only one tail. And those with two tails are only made 
use of before notes which cannot be deprived of any [sic] of their value; 
e.g. before two or more long notes, whether they are crotchets or minims, 
if they are at the same pitch. . . . These little notes with double tail are 
rendered very briefly in whatever manner they come to be taken, from 
below or from above; and they are played on the beat of the main note.' 



Observe that Quantz does not say that all short will 

be found written with the little semiquaver symbol; he merely 
any appoggiatura which is found so written is certain to be short. 
When he says "notes which cannot be deprived of any of their vaiue^ he 
is exaggerating; he means, no doubt, any measurable part of their 
value. The only appoggiatura which takes literally no value away from 
the main note, because it actually coincides with it, Is the simultaneous 
acciacatura (crashed appoggiatura), for which see Ch. XVII below. 

(201) C. P. E. Bach, Essay, Berlin, 1753, II, II, 13: 

'It is only natural that the Invariable short appoggiatura should most 
often appear before quick notes. It is written with one, two, three or 
more tails and played so fast that the ensuing note loses hardly any of its 
length. It also appears before long notes when a note is repeated [or] with 

[14:] 'When the appoggiaturas fill in leaps of a third, they are also taken 
short. But in an Adagio the feeling Is more expressive If they are taken as 
the first quavers of triplets and not as semiquavers. . . . The appoggia- 
turas before [actual] triplets are taken short to avoid obscuring the 
rhythm. . . . When the appoggiatura forms an octave with the bass it 
is taken short. . . . [15:] If a note rises a second and at once returns . . . 
a short appoggiatura may well ornament the middle note/ 

Ex. 51. C. P. E. Bach, short appoggiaturas: 


The above are the most interesting of C. P. E. Bach's numerous 
examples. The rule that appoggiaturas filling in leaps of a third are 
taken short must obviously be accepted with reservations; C. P. E. 
Bach himself gives many instances to the contrary, and it is only a 
certain type of passage which he has in mind here. His example is this 
(the interpretations are mine, but carry out his instructions). 

Ex. 52. C. P. E. Bach, Essay, Berlin, 1753, II, ii, 15, appoggiaturas 
filling in leaps of a third : 


It will once again be realised that the notation with tails suggestive 
of the timing is not even consistently applied by C. P. E. Bach (the last 



example should have double tails on his own principle) and was never 
in standard use. 

The term Invariable* applied to the short appoggiatura at the begin- 
ning of the above quotation also requires a warning. The length of 
short appoggiaturas varies within narrower limits than the length of 
long appoggiaturas, and in view of this C. P. E. Bach called the former 
'invariable' and the latter 'variable*. Yet he does not in practice treat 
the short appoggiatura as invariable; and, of course, it is not. Nor does 
the variability of a long appoggiatura extend indefinitely. It does not 
(as has sometimes been wrongly inferred) extend to the point of turning 
a long appoggiatura into a short one. C. P. E. Bach's terminology here, 
though understandable, is unfortunate. 

(b) The following is a particularly typical use of a short (semiquaver) 
appoggiatura in a context which would equally admit of a long (quaver) 
one. Notice how much better, however, the short appoggiatura carries 
on the rhythmic pattern of the first part of the bar. 

Ex. 53. F. W. Marpurg, Anleitung zum Clavierspielen, 2nd ed. 5 
Berlin, 1765, 1, ix, Tab. Ill, Exx. 43-44: 

(c) The following are four different ways of notating the standard 
short appoggiatura, in its characteristic position between notes a third 
apart. In other contexts, or even in contexts similar to this, (a), (c) and 
(d) might alternatively indicate long appoggiaturas. It rests with the 
performer to decide which fits best and makes the most expressive 
effect; but he should bear in mind the strong disposition towards the 
longer appoggiaturas throughout the eighteenth century. 

Ex. 54. F. W. Marpurg, Anleitung zum Clavier spielen, 2nd ed., 
Berlin, 1765, Table III, Figs. 11-13: 




(d) The post-Baroque use of/ to indicate a short appoggiatura began 
as a South German form of the ordinary semiquaver (E. and P. Badura- 
Skoda, Mozart-Interpretation, Vienna and Stuttgart, 1957, p. 83n). It 
was not a separate sign till the nineteenth century. 




(202) Joachim Quantz, Essay* Berlin, 1752, VIII, 12: 

'It is not enough to know how to perform the appoggiaturas according 
to their nature and difference, when they are marked ; it is also necessary 
to know how to put them in suitably when they are not written. Here is a 
rule which can be used for understanding them. When after one or more 
short notes on the down or up beat of the bar a long note comes, and 
remains in consonant harmony; an appoggiatura must be put before the 
long note, to keep the melody continuously agreeable ; the preceding note 
will show whether the appoggiatura should be taken from above or below.' 

In reality the harmony will gain even more than the melody in the 
context here envisaged. Such a context also implies that the appoggia- 
tura should be long. 

(203) C. P. E. Bach, Essay, Berlin, 1753, H, ii, 8: 

'Because the sign for the appoggiatura with that of the trill is almost 
the only one familiar everywhere, w r e furthermore generally find it 
written down. Yet since this cannot always be depended on, it is necessary 
to decide on the proper contexts of the variable [i.e. long] appoggiatura, so 
far as this can be done. 

[9:] 'The variable [long] appoggiatura in duple time frequently appears 
either on the down beat (Fig. 70, Ex. a) or the up beat (K) ; but in triple 
time only on the down beat (Fig. 71) and that always before a compara- 
tively long note. It is also found before cadential trills (Fig. 72, Ex. a) ; 
half-closes (6), caesuras (r), fermatas (d) and closing notes with (e) or 
without (/) a previous trill. We see from Ex. e that the ascending 
appoggiatura is better than the descending after a trill; thus the instance 
(g) is weak. Slow dotted notes also bear the variable [long] appoggiatura 
(A). When such notes have tails, the tempo must be such as to be suitable/ 

Ex. 55. C. P. E. Bach, contexts inviting unwritten appoggiaturas: 

It will be realised that in the above examples, the little notes represent 
the appoggiaturas which C. P. E. Bach suggests should be put in by 
the performer in cases where no indication of them appears in the 
written text. 





(a) In rccitati\e, and to a lesser extent in arioso passages as well, 
very conventions applied concerning the introduction 

of by the performer. 

J. A. in F. W. Marpurg's Kritische-Briefe, Berlin, for 

1760-62, letter 109, p. 352: 

"Regarding the of the feminine cadence with the falling 

fourth, we must observe that the last two notes are written by some 
differently from what is sung, e.g. : 

Ex. 56. Seheibe, misleading notation of feminine endings: 

"" i. "^fritter; :"] :npLu k cof 

"This notation is undoubtedly regrettable because in the nature of 
things we should not write differently from what is sung, and because 
many untutored singers can be led astray by this, especially in the middle 

of a recitative.' 

(b) By a feminine cadence is meant a close (whether passing or final) 
in which the last syllable falls not on the beat (which is the masculine 
ending) but off the beat. It will be noticed that Scheibe speaks of this 
not as a matter of optional ornaments, but as a definite misnotation 
of which the performer has to know, and apply, the intended interpre- 
tation. I am quite certain that this is not a decision in which any 
option arises. Most conductors and singers are inconsistent or negligent 
in their application of this convention, although very few are altogether 
without knowledge of it. Some conductors deliberately resist it on the 
grounds of wanting only what the composer wrote. As so often in 
baroque music, this gives them what the composer wrote at the expense 
of what he meant by what he wrote. There are few rules so unambiguous 
as this rule concerning appoggiaturas in recitative, and its consistent 
application is very strongly to be recommended. 

Ex. 57. G. P. Telemann, Cantatas, Leipzig, 1725, preface, feminine 
and masculine endings in recitative: 


(c) Of the above, the fir>t two are examples of feminine endings, the 

last two of masculine endings. Telemann treats both alike, makes, 
no suggestion of any possible exceptions in either case. J. K Agricola's 

German translation (Berlin, 1757) of Tosfs includes his. 

examples of masculine endings, all likewise given normal appoggiaturas. 
On the other hand, Scheibe's rule was given specifically for feminine 
endings, and he does not mention masculine ones at all. Thus we are not 
categorically precluded by his evidence from making an exception in a 
masculine ending where there seems a strong musical reason for doing 
so. Both the following endings from J. S. Bach's St. Matthew Passion are 
masculine: (a) implies the normal appoggiatura; but the phrase at (b), 
which is one of peculiar forthright ness and power, sounds stronger if 
the second A balances the first A exactly, without being complicated 
by an appoggiatura. We have no evidence suggesting that an exception 
would, even so, have been made at the time; but it is exactly the sort 
of rare case where we might Just possibly make an exception now. 
And, of course, outside recitative or near-recitative the situation is 
much more fluid. 

Ex. 58. J. S. Bach, masculine endings (St. Matthew Passion* recit. 
No. 4 in Novello's ed., last two bars): 

and lull him But they sale!, smdT lall hlxnt But they sal J 

Ex, 59: the same (recit. No. 32, bar 26): 

as written ami ?paibrrr.ed (b) 

that is Hr; hold him. fart fast 

Ex. 60. G. P. Telemann, Cantatas, Leipzig, 1725, preface, mixed 
masculine and feminine endings: 



Id) That this convention mas itself subject to further ornamentation 

by the performer is suggested by the following written-out example, 
it should only be imitated with considerable discretion and 
restraint, in carefully selected contexts. 

Ex. 61. J. S. Bach, Christmas Oratorio (B.-G., V, 2, p. 236, bar 9), 

written-out ornament : 


kom-me u:i2 t* an* be - - ta 


(a) Both the long and the short appoggiattira continued to be taught 
by the treatises of the late eighteenth century. The long appoggiatura 
tended to grow not shorter, but longer still. 

(205) D. G. Turk, Klavierschule, Leipzig and Halle, 1789, nine rules 

for appoggiaturas, including: 

Ex. 62. Turk, very long appoggiaturas: 

'Every long appoggiatura to be played with more emphasis then the 

ensuing note.' 

(206) G. F. Wolf, Unterricht, 3rd ed., Halle, 1789, p. 69, includes: 


'[Short appoggiaturas appear only before short notes, when] the main 

note itself loses almost nothing at all,' 

Hoyle's Dictionary of Music, 1770, pirates Genuniani's description 
of the long and short appoggiatura for which see (189) in 6 above 
almost word for word. 

Ex. 64. J. C. F. Rellstab, C. P. E. Bach's Anfangsgrunde mil einer 
Anleitung, Berlin, 1790, VIII, very long appoggiatura: 



(20?) Thomas Bush), Complete Dictionary nf \fnsk\ London, [1786], 

5.1*. "Appoggiatura or Leaning \ote*: 

The appoggiatura not being always in consonance with the 

other parts, to avoid a visible breach of the laws of harmony, it is generally 

written in a small note. 

(b) Cramer, Pleyel, Czerny and Hummel in the early 

century taught similar rules for the appoggiatura. J. D. Andersch's 
Musikalisches Woerterbuch (1829) makes the standard distinction 
between long and short appoggiaturas. 

(208) Louis Spohr, Violinschule, 1832, etc., rules for appoggiaturas: 

If the appoggiatura stands before a note which can be divided into 
equal parts, it obtains the half of its value. . . . Before a note with a dot 
it obtains the value of the note, which then begins only at the dot. . . . 
Where there are two dots the appoggiatura obtains the value of the note 
and this then begins with the first dot.' 

(c) The traditional appoggiaturas continued to be taught, though 
with some elements of confusion, throughout the nineteenth century. 

(209) F. de Courcy, Art of Singing, London, c. 1868: 

'[Appoggiaturas are] generally half the succeeding or principle note.' 

(d) In the twentieth century, the more traditional schools of Italian- 
ate singing retained the standard long and short appoggiaturas, and 
there are still a few circles in which they are not forgotten. It must be 
remembered that such a classic as Manuel P. R. Garcia's Traitt complet 
de rart du chant (Paris, 1847) is in regular use by a few teachers who 
retain the traditional Italian methods at the present time: I consulted 
an edition of 1911 and found it very sound on ornaments. But the 
majority of modern teachers and their pupils are no longer working in 
the old tradition, and confusion as to the length and the impromptu 
introduction of appoggiaturas is now general. 

(e) It will be seen from the above that for most music down to about 
the middle of the nineteenth century, and for some music after that, 
the principles which governed both the long and the short appoggiatura 
in the baroque period apply with undiminished force. Thus in Mozart's 
operas, for example, many appoggiaturas (particularly long appoggia- 
turas) are so plainly implied that their neglect amounts to a disregard 
of the composer's intention. Charles Mackerass ('Sense about the 
Appoggiatura', Opera, Oct. 1963) quotes orchestral parts having 
written-out appoggiaturas in unison with vocal parts not showing (but 
certainly meaning) these appoggiaturas. A few typical situations, and 
my suggested solutions for them, are as follows. 



Ex. 65. Mozart, Cost Fan Tutte (recit. and aria No. i 1 in NoveUo's 

>>hr my tin - p 

Wkfc wiE co -scfcxne - joteme 


BottiKBlhqrnaghtbe slam fci& 

AKD ft i K K ^ Pgfam: K K i . TiA TJ 

\\Tiatarey0u saying saying I on -Jy ay what's true true 

A typical appoggiatura problem appears in the slow movement of 
Beethoven's posthumous quartet, Op. 127. In view of Beethoven's 
admiration for C. P. E. Bach's Essay, which according to Czerny's 
autobiographical notes (cited L. Nohl, Beethoven Depicted by his 
Contemporaries, 1877, tr. E. Hill, 1880, Ch. VII) Beethoven recom- 
mended to his pupils, and in view of the generally continued currency 
of the rales for appoggiaturas as given by C. P. E. Bach, there can be 
no real doubt that the appoggiaturas in this movement are not short 
(as played even by so traditional a quartet as the Flonzaley, which is as 
far back as my memory goes) but long. They should be performed as 
dotted crotchets. The effect sounds somewhat revolutionary at first, but 
when its novelty has worn off becomes most convincing, both melodi- 
cally, and still more harmonically. It is, in fact, what Beethoven intended. 

In Mozart's Piano Concerto K. 459, second movement, the 
appoggiaturas in the orchestral parts at bar 67, etc., are nowadays, but 
incorrectly, taken short (and often, still more incorrectly, before the 
beat as well). They should be long (i.e. half the length of the quavers on 
which they stand) as is proved by their having to match the piano part 
at bar 71 , etc. This again is a typical case of modern misunderstanding. 
For one such case where proof positive is afforded (in this case by the 
precisely notated piano part) there must be many hundreds of others 
where there is no specific proof, but where the standard rules make the 
correct interpretation virtually certain. 

As an instance of a modern on-the-beat short appoggiatura, it may 
be mentioned that Mahler, whose use of very numerous, very long 
appoggiaturas is a feature of his style, but who writes these long 
appoggiaturas out, also writes a few ornamental appoggiaturas in little 
notes. These were taken by Bruno Walter, who directly inherited his 
tradition, definitely short but with a lingering expression, and firmly on 
the beat. To anticipate the beat with them would spoil the whole effect. 



A (ii): The Double Appoggiatura 

French; Port-de-volx double 
German: Ansch!ag f Doppelvorschlag 


The clearest name for this ornament is that used by Ape! in his Harvard 
Dictionary of Music (1944): disjunct double appoggiatura. This dis- 
tinguishes it from the Slide, well called by Apel: conjunct double 
appoggiatura. But even these names are not ideal, since double, triple 
appoggiatura, etc., have also the meaning of two, three or more appog- 
giaturas going on in different parts of a chord at the same time. 

The reason for calling the ornament here under consideration a 
double appoggiatura at all was that two notes behave in some respects 
like the one note of which the appoggiatura proper consists. These two 
notes are always disjunct with each other. The second is always con- 
junct with and nearly always superior to the ensuing main note. 

This ornament is not of widespread importance nor much to be used 
outside the galant school. 


(210) Joachim Quantz, Essay, Berlin, 1752, XIII, 41: 

'[The double appoggiatura] must be joined very quickly, but softly, to 
the [main] note. The [main] note must be slightly louder than the accessory 

Ex, 66. Quantz, double appoggiaturas: 


Ex. 67. F. W. Marpurg, Principes du clavecin, Berlin, 1756, double 
appoggiaturas (interpretations below each): 


Notice the curious effect of the dotted double appoggiaturas In the 
last but one of the above examples. C. P. E. Bach (Essay, 1753, 1, ii, 20) 
confirms this dotted form, but not in rapid passages. He gives three 
general rales. 

(211) C. P. E. Bach, Essay, Berlin, 1753, II, vi, 3: 

4 [The accessory notes] comprising a double appoggiatura are played 
more softly than the main note. [7:] the [main] note yields as much of its 
length as the ornament requires. [10:] all three notes are slurred.' 

The second of these rules shows that the double appoggiatura is 
taken on (and not before) the beat, in spite of its far from appoggiatura- 
like behaviour in being softer than its main note. Its function is more 
rhythmic than melodic or harmonic; it is not itself accented, but 
displaces the accent of its main note to an irregular position just after 
the beat. 

But in view of the technical and musical difficulty of carrying out 
these instructions literally, there is some reason to suppose that it may 
also have been used as a normal rhythmic ornament taking the accent 
as well as the beat. 


When encountered in composers of the Romantic school, such as 
Chopin, the double appoggiatura is to be taken on the same principles 
as in the baroque period. 



A(iii): The Slide 

English: Elevation 

Double backfall 



French: CouM 

Coule sur tine tierce 

German: Schlelfer 



Latin: Superjectio 


The name slide Is admirably fitted to this ornament, which slides 
rapidly and smoothly through its two conjunct accessory notes to its 
main note whence its alternative description as a conjunct double 
appoggiatura. It is a very widespread ornament in different ages and 
localities, and is of considerable value and importance in all stages of 
baroque music. 


(212) G. Caccini, Nuove musiche., Florence, 1602, tr. Playford, Intro- 
duction, London, 1654, ed. of 1674, p. 42: 

'There are some therefore that in the Tuning of the first Note, Tune it a 
Third under ... it agrees not in many cords, although in such places as 
it may be used, it is now so ordinary, that instead of being a Grace (because 
some stay too long in the third Note under, whereas it should be but 
lightly touched) it is rather tedious to the Ear.' 

No mention is made here of the intervening note, but in other 
respects the description fits the slide, and this seems to be intended. 
Bovicelli (Regole, 1594) and Praetorius (Syntagma, III., 1619) include 
the ascending slide, in the form evidently disliked by Caccini, in which 
its first note is dotted. Praetorius also shows it undotted. 

Being undoubtedly a very common ornament at the turn of the 
sixteenth and the seventeenth centuries, the slide has been suggested 
by Thurston Dart (The Interpretation of Music, London, 1954, p. 120) 



for the ornament of the English Virglnaiists. This sugges- 

tion from the fact that a slanting stroke (though not 

normally across the stem) Is found later In the seventeenth century for 
ascending notes forming or resembling slides, among other meanings. 
In many contexts, a quick slide makes a musically convincing solution 
for this single-stroke Virginalists' ornament; on the other hand, there 
are other contexts where it does not serve so well, if at all. It may 
therefore be tetter to confine ourselves to regarding the slide as one of 
the two or three possible solutions at a choice of which this sign hints, 
perhaps in common with the double-stroke so often associated with it 
or substituted for it: the other possibilities being a rapid lower mordent 
and a (usually brief) half-trill. Occasionally a full trill with turned 
termination may be the right solution. 

Ex. 68. Playford, Introduction, London, 1654 (1660 on), descend- 
ing and ascending slides: 



A Double Backfall 
t ^ 


Simpson (Dwision-Violist, 1659) agrees; Mace (Mustek's Monument, 
1676) gives an ascending slide under the name of whole-fall, and its 
sign as x. Purcell or his editor of 1696 (Lessons) gives an ascending 
slide under the name of slur, probably as a translation of the French 

Ex. 69. Chambonni&res, Pieces de clavecin, Paris, 1670, slide: 

Ex, 70. D'Anglebert, Pieces de clavecin, Paris, 1689, slides: 

m t ii r iff 

^Coulesur Autre Sur 2 not 

ts. Autre 

J. G. Walther shows slides taken, not on the beat, but before it, and 
though he was not a leading authority, had presumably heard them so. 
When slurred to the following main note, they have as much claim to 
be viewed as genuine slides irregular in their behaviour as passing 



appoggiaturas have to be viewed as a 

claim, but a recognisable one, provided it is 

usages are complete departures from the standard practice, (J. D. 
Heinichen General-Bass, 1728, I, vi, 9- also describes a 
(Schleiffung) in a similar position; but his execution merely 

passing notes between beats, there being no slurring to the following 
note.) The slide always sounds stronger, and In my opinion tetter, on 
the beat. 

Ex. 71. J. G. Walther, < Praecepta% 1708, slides anticipating the 





1 . L. J 
3 according to W: 




-=** 9T 4 

J. S. Bach sometimes uses the sign shown in the above example, but 
intends the slide to come in Its correct position, i.e. on the teat, as 
other examples which he has written out In full make clear. 

(213) Joachim Quantz, Essay, Berlin, 1752, XVII, ii, 21: 

'When, in a slow movement, one finds added small quavers, the first 
of which is dotted, they take the time of the succeeding main note, and 
the main note takes only the time of the dot. They must be played 

'[23: undotted slide] belongs to the French style rather than the Italian 
[and] must not be played so slowly [but] with rapidity.' 

The first example next to follow is an extreme case; the remainder 
are normal variants. 

Ex. 72. C. P. E. Bach, Essay, Berlin, 1753, II, vii, very extended 
slide; normal slides: 


ritten performed 

7 r ' 

Ex. 73: 


Ex. 74. F. W. Marpurg, Kunst das Clavier :u spielen, Berlin, 1750, 

flr ifc ft 


Hummel, a modernist in matters of ornamentation, nevertheless 
shows (1828) slides on the beat in the traditional baroque fashion. 
This appears to be the standard interpretation for the nineteenth 

The slides with which Beethoven's last quartet, Op. 135, opens 
should take both the accent and the beat. The figure is greatly weakened 
if they are allowed to come before the beat. The same applies, of 
course, throughout the movement. The case is typical 



A (iv): The Acciaccatura 

English: Acciaccatura 

Crashed appoggiatura 

Italian: Acciaccatura 


French: PInce etouffe 
German: Zusammenschlag 


(a) The Italian (and borrowed English) name acciaccatura means a 
'crashed stroke'; the German equivalent means a "together stroke*. 
Both are descriptive of the manner of performing the primary orna- 
ment, which is to strike simultaneously the main note and an accessory 
note a semitone or tone below, but to release the accessory as soon as 

The French name means a 'smothered mordent*, which indicates the 
lines on which some authorities, such as Marpurg and C. P. E. Bach, 
explained the ornament. 

(b) The name was also extended secondarily by Geminianl and others 
to indicate a little ornamental note foreign to the harmony and 
interpolated momentarily in chords, particularly if arpeggiated. 

(c) We may distinguish the first of these two forms (the one now 
ordinarily associated with the name) as the simultaneous acciaccatura; 
the second as the passing acciaccatura. 

(d) It will be evident that both forms of the acciaccatura are feasible 
only on keyboard instruments. Both are valuable, but the second was of 
much more importance than the first, being a regular impromptu 
resource of keyboard players though not very frequently notated. 


(214) Francesco Geminiani, Treatise of Good Taste, London, 1749: 

*Tatto [simultaneous acciaccatura] which has a very great and singular 
Effect in Harmony, and which is perform'd by touching the key lightly, 
and quitting it with such a Spring as if it was Fire.' 


(215) C. P. E. Berlin, 1753, II, v, 3: 

*. . .an unusual manner of executing a very short mordent [simultaneous 

acciaccatura). Of the two notes struck together, only the upper one is 
held, the lower one being at once released. . . . It is only used abruptly, 
that is to say in detached passages/ 

Ex. 75. C. P. E. Bach, Essay, Berlin, 1753, II, v, 3, simultaneous 

acciaccatura : 

(216) F. W. Marpurg, Principes du Clavecin, Berlin, 1756, XIX, 2: 

4 Instead of playing the two keys successively, they are often struck both 
together, but the accessory note is held for only half its value, so that the 
main note may be heard by itself afterwards. . . . This kind of mordent 
is called Pinct Itouffl, in Italian Acciaccatura, and it is frequently used in the 
bass. In transitions from soft to loud it can be effectively used to emphasise 
the harmony. 5 


(217) Fr. Geminiani, Treatise of Good Taste, London, 1749: 

"The Acciaccatura is a Composition of such Chords as are dissonant with 
respect to the fundamental Laws of Harmony ; yet when disposed in their 
proper place produce that very Effect which it might be expected they 
would destroy.' 

Ex. 76. C. P. E, Bach, Essay, Berlin, 1753, III, 26, 'an arpeggio 
with an acciaccatura^ : 

written performed 

The acciaccatura in the above example, giving as it does two notes 
a third apart which are held down while the dissonant note between is 
released, has precisely the form of one of the varieties of the slide; and 
it is here notated with the same sign. This sign is also used for an 
acciaccatura which does not join notes a third apart, and is therefore 
different from a slide. 

Ex. 76a. J. S. Bach, Partita VI, Sarabande, acciaccaturas ornament- 
ally notated by slanting strokes and by a little note: 



(a) written ^^^^^^ ^ natfrlttsr. 

c^J=irr=arr^^ ^4r"3"!^ 


1 i 


The notation used by J. S. Bach was as at (a). The alternative notation 
shown at (b) Illustrates a still commoner method of writing down the 
passing acciaccaturas with which a chord may be enriched, particularly 
if, as here, it is full enough to be well arpeggiated. Domenico Scarlatti 
was especially fond of introducing thick chords which look on paper 
impossibly discordant for the period, because they Include their passing 
acciaccaturas fully written out in the manner of (b) above. However 
slightly such chords are to be arpeggiated, there is always time to raise 
the interpolated discords while holding down the essential notes of the 
chord; the distinction may be very momentary indeed, but the ear will 
pick it out immediately, hearing the acciaccaturas as fleeting disturb- 
ances of an intelligible harmony. The effect can be extraordinarily 
subtle, rich and beautiful; and it can be very freely introduced in 
suitable contexts, when not written (as, of course, it usually is not). 
Indeed, in those passages of accompaniment which are best served by 
freely arpeggiated chords (figured arpeggios), the passing acciaccatura 
is of almost continuous utility. This subject will be treated partly in 
Ch. XXIV, 1, as a branch of ornament, and partly in Ch. XXX as a 
branch of accompaniment. In solo passages of a certain type, particu- 
larly the unmeasured preludes of the French school, the passing 
acciaccatura is equally of value. 

(217a) Vincenzo Manfredini, Regole armoniche, Venice, 1775, p. 62: 

'These Acciaccature make a better effect when the Chords are executed 
in Arpeggio form, as one is accustomed to do when accompanying 


(a) The crushed acciaccatura remained in limited but accepted use in 
the late eighteenth century, and also, since it is very effective on the 
piano, in the nineteenth century. 



In the twentieth century it is not entire!)' obsolete, being correctly 
by some; but by the majority it has become confused with the 
ordinary short appoggiatura whether correctly performed on the beat, 
or (as more often) incorrectly before the beat. Even if the short appog- 
giatura is correctly performed, however, it is a regrettable substitution, 
since the true acciaccatura remains a splendidly virile ornament whose 
displacement is to be deplored. 

(b) The passing acciaccatura did not remain a standard nineteenth 
century ornament. In the twentieth century, effects of fleeting dissonance 
somewhat resembling it are found written out in certain styles, but the 
device of holding the essential notes while releasing the dissonant inter- 
polators is not so typical now that our familiarity with such inessential 
dissonances makes it less necessary to remove them as quickly as 
possible from the harmony. 



A (v): The Passing Appoggiatura 


German: Nachschlag 

Durchgehender Vorschlag 


This name is the translation of the second and more explicit of the 
two German names above, which means literally: through-going before- 
stroke. As that indicates, it was regarded as an appoggiatura, not as a 
mere passing note. Its claim to be so regarded rests, however, only on 
the opinion of a few mid-eighteenth century authorities, and on a single 
fact: it was slurred to the following and not to the previous note, which 
would have been its natural behaviour as a passing note. 

The fashion for this ornament appears to have been of no very great 
duration, and it is important chiefly because it looks in notation like 
a normal appoggiatura, so that it is not very obvious at first sight 
which is intended. 


(218) Joachim Quantz, Essay, Berlin, 1752, VIII, 5: 

'There are two sorts of appoggiaturas. One ought to be introduced, 
like accented notes, on the down-beat; the others like unaccented notes, 
on the up-beat of the bar. The first could be called striking [Ger. ed: 
anschlagende; Fr. ed. : jrappant], and the other passing [durchgehende; 
passagers] appoggiaturas. 

[6:] 'Passing appoggiaturas are found when several notes of the same 
value descend by leaps of a third [Ex. (a)]. They are rendered in practice 
as [Ex. (b)]. It is necessary to sustain the dots, and give a stroke of the 
tongue [i.e. on the flute, concerning which Quantz is ostensibly writing] 
on the notes where the slur begins, that is to say on the second, the fourth 
and the sixth note. This sort must not be confused with the notes where 
there is a dot after the second, and which yield almost the same melody 
[Ex. (c)]. In this latter figure the second, the fourth, and the subsequent 
short notes come [as normal accented appoggiaturas] on the beat of the 
bar, as dissonances against the bass; moreover they are rendered with 
boldness and liveliness. The appoggiaturas with which we are here con- 
cerned, on the contrary, require a flattering expression. If then, the little 
notes at [Ex. (a)] were taken long and given a stroke of the tongue on the 



beat which follows, the melody would be completely changed [from] the 
French style of performing, from which these appoggiaturas are derived 

[7:] 'Striking appoggiaturas, or those which come on the beat of the 
bar, are found before a long note on the down-beat, following a short one 
on the up-beat [and are] held for half the following main note.' 

Ex. 77. Joachim Quantz, Essay, Berlin, 1752, VIII, 5: 

^ j^ i K . i 





in 1 r IT IT 


not to be confused with: 


which Is one of the ways of playing:: 



Except that Quantz appears in the last paragraph to be contrasting 
the passing appoggiatura with the long appoggiatura, and in the 
previous paragraph with the short appoggiatura, the above is clear; 
but though other authorities agree that the passing appoggiatura 
normally fills in a melodic leap of a third, they do not confine it to 

(219) Leopold Mozart, Violinschuk, Augsburg, 1756, IX, 17: 

'Passing appoggiaturas ... do not belong to the time of the main note 
on to which they fall but must be taken from the time of the note 
before. . . . The semiquaver is taken quite smoothly and quietly, the 
accent always coining on the quaver.' 

Ex. 78. Leopold Mozart, Violinschule, 1756, IX, 17, passing 
appoggiaturas : 

'Without embellishmentTiius it could be written? 
down up down 


%ut they arc played thus, 
and are Better written so* 

Ex. 79. Jean-Jacques Rousseau, Dictionnaire de musique, Paris, 
1768, Plate B, passing appoggiatura: 


$ ' 

fl performed 





(220) F. W. Marpurg, Anleitung, 2nd ed., Berlin, 1765, I, ix, 4: 

'All appogglaturas ... in whatever progression they occur, must fall 
exactly on the beat. Therefore it is wrong if [Ex, (a) is] performed as at 
[Ex. (b)] or even as at [Ex. (c)]. They must be played as at Ex. (d).' 

Ex. 80. F. W. Marpurg, Anleitung, 2nd ed., Berlin, 1765, Tab. HI, 
examples of passing appoggiaturas condemned : 

fa) \vritten (b) [wont*] (c) [wrung] (dj [right] 

A . -^ c-T cta 



B (i): Tremolo and B (ii): Vibrato 


Old English: Organ shake 


Italian: Tremolo 



Mordente fresco 
French: Balancement 



Verre Casse 
German: Bebung 


There is great confusion of terms here. But the clearest terminology is 
to reserve the word tremolo for a fluctuation of intensity, often but not 
necessarily amounting to a reiteration of the note; and the word 
vibrato for a fluctuation of pitch not amounting to a change of note. 

When two notes a tone or semitone apart alternate, we speak of a 
trill in the normal modern sense. When the notes are more than a tone 
apart, we again speak of a tremolo or tremolando, whether these notes 
are reiterated simultaneously or alternately. 

But at the beginning of the baroque period, the Italian word trillo, 
and its abbreviations /, tr, tri, etc., usually meant not trill, in the 
modern sense, but what is here called tremolo (on one note at a time). 

All the above were regarded as specific ornaments in the baroque 


(a) It is tremolo, not vibrato, with which a good singer brings his 
tone to life. 

(b) The following refers to a particular development of the tremolo, 
carried to the point at which the fluctuations of intensity are slow 
enough, and just but only just distinct enough, to sound like a reitera- 
tion of the note. In this form, the tremolo is treated as an ornament: 
i.e. the early baroque Italian vocal trillo. 



(221) Playford, Introd., London, 1654 (1664 on), cited from ed. of 
1674, on Caccini's Preface, Nuove Musiche, Florence, 1602: 

'The Trill [here meaning tremolo] is by a beating in the Throat . . . 
rather the shaking of the Ovula or Pallate on the Throat, in one sound, 
upon a Note. ... I have heard of some that have attained it by this 
manner, in singing a plain song, of 6 Notes up and 6 down, they have in 
the midst of every Note beat or shaked with their finger upon their 
throat, which by often practice came to do the same Notes exactly with- 
out. . . . The Trill, or Shake of the Voice, being the most usual Grace, is 
made in Closes, Cadences, and other places, where by a long note an 
Exclamation or Passion is expressed, there the Trill is made in the latter 
part of any such Note; but most usually upon binding Notes in Cadences 
and Closes, and on that Note that precedes the closing Note. Those who 
once attain to the perfect use of the Trill, other Graces will become easie/ 

This accords with a vaguer account given by Praetorius; his illus- 
tration does not add up, but appears to be meant as follows. 

Ex. 81. M. Praetorius, Syntagma, III, Wolfenbiittel, 1619, p. 237, 

As Praetorius rightly adds, 'this kind is found in Claudio Monte- 
verdi'. The ornament was correctly described by Michel L'Affilard in 
his Principes of c. 1694, and by Brossard, who gives the following 
important information showing that the above example stands for a 
continuous acceleration, and not literally, as written, for a graded one. 

(222) Brossard, Dictionaire, Paris, 1703, s.v. 'Trillo': 

'TV . . .is very often, in Italian music, the sign that one must beat 
several times on the same note, at first somewhat slowly, then ending 
with as much lightness and rapidity as the throat can make . . . [but] our 
example can give only a very crude idea of it, compared with the quick- 
ness with which it can be done/ 

Brossard's example suggests that the acceleration might be not 
regular, but with a slight hesitation of rhythm. This, though not im- 
possible musically, is presumably a mistake, since subsequent editions 
have the example corrected to a regular acceleration. 

Ex. 82. Brossard, Dictionaire, s.v. 'Trillo', 'the true Italian 
"Trillo 5 ": 

thus J ^^ 




But Tosi in his Opinion! of 1723 disapproved of the ornament, which 
he described under the name of mordente fresco. It is unlikely that the 
ornament survived much longer. 

(c) There Is no reason to exclude from baroque music a moderate 
use of the ordinary tremolo not as an ornament but as the means by 
which singing tone is kept alive and vibrant, just as string tone is kept 
alive by the vibrato, for which see 4 below. 

(223) Praetorius, Syntagma, III, Wolfenbiittel, 1619, p. 231: 

A singer must have a fine, pleasing, trembling and shaking (zittern 
und bebende) voice, yet not used as in some schools, but with especial 


(a) Tremolo in the sense of rapidly reiterated notes was written for 
the violins by Monteverdi in his Combattimento di Tancredi e Clorinda 
(1624): probably meant measured, i.e. as written, not as a free tremo- 
lando. This was no novelty in music, but the peculiarly expressive 
purpose to which the device is there put may well have been. Un- 
measured tremolando is, however, a possible interpretation. 

(224) Christopher Simpson, Division- Violist, London, 1659, p. 10: 

'Some also affect a Shake or Tremble with the Bow, like the Shaking- 
Stop of an Organ, but the frequent use thereof is not (in my opinion) 
much, commendable.' 

This undoubtedly free (unmeasured) tremolando is at least a feasible 
interpretation of the much disputed wavy lines in the famous frost 
scene in PurcelTs King Arthur. 

(b) Instruments plucked with fingers or plectrum are amenable to 
thrumming, which is a species of tremolo ; but this was apparently not 
part of classical lute technique. 

(c) The clavichord has a species of vibrato which if intensified pro- 
duces a genuine tremolo. The general names for this technique are It. 
tremolo, Fr. balancement, and Ger. Bebung. The second aspect is some- 
times distinguished as Tragen der Tone. But the two things are so 
closely allied that their names also tend to be confused. 


Old English: Close shake 


Latin: Tremor Pressus 

Italian: Vibrato 



French: Aspiration 








Tremblement mineur 

Verre casse 
German: Bebung 



From the middle of the baroque period until the generation of 
Kreisler, there was controversy whether the instrumental vibrato 
should be used (in the modem way) more or less continuously as a 
means of enlivening the tone, or intermittently as a specific ornament. 
But that it was used throughout the baroque period, there can be no 
doubt at all; nor have we any reason to suppose that it was then a 
novelty. Both Martin Agricola (Musica Instrumentalis Deudsch, ed. of 
1545, 42-3) and Ganassi (Regola Rubertina, 1542, Ch. II) mention 
a "trembling" of the fingers which is almost certainly vibrato. 

(224a) M. Mersenne, Harmonie Universelle, Paris, 1636-37, tr. R. E. 
Chapman, The Hague, 1957, Book II, sect, on Lute ornaments, p. 24: 

'The tone of the violin is the most ravishing [when the players] sweeten 
it ... by certain tremblings [here meaning vibrato] which delight the 

[p. 109:] 'The verre casse [here meaning vibrato] is not used so much 
now [on the lute] as it was in the past [partly in reaction] because the older 
ones used it almost all the time. But [it cannot be dispensed with and] 
must be used in moderation . . . the left hand must swing with great 
violence [the thumb being free of the neck]. 

Ex. 83. Playford, Introduction, London, 1654 (1660 on), Table of 
Ornaments, vibrato: 

(225) Christopher Simpson, Division-Violist, London, 1659, sect. 16: 

'Close-shake is that when we shake the Finger as close and near the 
sounding Note as possible may be, touching the String with the Shaking 
finger so softly and nicely that it make no variation of tone. This may be 
used where no other Grace is concerned.' 

Two fingers are used for this form of vibrato. By 'no variation of 
tone' Simpson evidently means (as an American though not an English 



writer would now) "no change of note'. There is, however, a distinct 
fluctuation of pitch, and the result is a vibrato rather more prominent 
than usual. 

(226) Thomas Mace, Mustek's Monument, London, 1676, p. 109: 

'The Sting [vibrato], is another very Neat, and Pritty Grace; (But not 
Modish [on the lute] in These Days) . . . first strike your Note, and so 
soon as It is struck, hold your Finger (but not too Hard) stopt upon the Place, 
(letting your Thumb loose) and wave your Hand (Exactly) downwards, and 
upwards, several Times, from the Nut, to the Bridge; by which Motion, your 
Finger will draw, or stretch the String a little upwards, and downwards, so, as 
to make the Sound seem to Swell! 

This too results in a fluctuation of pitch rather than of intensity, 
and is a normal (one finger) vibrato. 

(227) Jean Rousseau, Traite de la Viok, Paris, 1687, pp. 100-101 : 

'The Batement [here meaning the extreme form of vibrato] is made 
when two fingers being pressed one against the other, the one is held on 
the string, and the next strikes it very lightly . . . [it] imitates a certain 
sweet agitation of the Voice ... is used in all contexts where the length 
of the Note permits, and should last as long as the note. 

'The Langueur [here meaning the normal vibrato] is made by varying 
the finger on the Fret. It is ordinarily used when it is necessary to take a 
Note with the little finger, and time permits; it should last as long as the 
Note. This ornament is to replace the Batement which is unavailable when 
the little finger is held down.' 

Marin Marais (Pieces de viole, Paris, 1696) gives Pince ouflattement 
for the two-finger vibrato and Plainte for the one-finger vibrato. In his 
music the sign for the former is a thin, horizontal wavy line, for the 
latter a similar line but vertical; the absence of either sign does not, of 
course, preclude vibrato he merely marks them where he wants to 
make particularly sure of them. (But Jean Rousseau means by Plainte 
a portamento?) 

(228) Fr. Geminiani, Art of Playing on the Violin, London, 1751, 
p. 8: 

'Of the Close SHAKE . . . you must press the Finger strongly upon 
the String of the Instrument, and move the Wrist in and out slowly and 
equally. When it is long continued, swelling the sound by degrees, 
drawing the bow nearer to the bridge, and ending it very strong, it may 
express majesty, dignity, etc. But making it shorter, lower, and softer, it 
may denote affliction, fear, etc., and when it is made on short Notes, it 
only contributes to make their Sound more agreeable and for this 
Reason it should be made use of as often as possible.' 

(229) Leopold Mozart, Violimchule, Augsburg, 1756, XI, iff.: 



*The Tremolo [here meaning vibrato] is an adornment which arises 
from Nature herself ... if we strike a slack string or a bell sharply, we 
hear after the stroke a certain undulation. . . . Take pains to imitate this 
natural quivering on the violin, when the finger is pressed strongly on the 
string, and one makes a small movement with the whole hand , . . 
forward and backward . . . 

'Now because the tremolo is not purely on one note but sounds undu- 
lating, so it would be a mistake to give every note the tremolo. There 
are performers who tremble [make vibrato] on every note without 
exception as if they had the palsy . . . 

'There is also a slow, an increasing, and a rapid undulation/ 

Whether or not vibrato 'arises from Nature herself', it certainly arises 
from the nature of bowed string instruments. It should be used with 
sufficient restraint to keep it in style; but it should certainly be used. 
String tone can sound very dead without it. (Like so much about 
ornaments in Leopold Mozart, the above is lifted, without acknowledge- 
ment, from Tartinfs MS. treatise on ornamentation see Bibl.) 

(c) Vibrato is a regular part of clavichord technique. 

(230) C. P. E. Bach, Essay, Berlin, 1753, III, 20: 

*A long, expressive note [on the clavichord] may be performed with a 
vibrato. The finger holds down the key and rocks it, so to speak/ 

(231) Dr. Ch. Burney, Present State of Music in Germany, London, 
1773, II, 268 (of C. P. E. Bach): 

*In the pathetic and slow movements, whenever he had a long note to 
express, he absolutely contrived to produce, from his instruments, a cry of 
sorrow and complaint, such as can only be effected upon the clavichord, 
and perhaps by himself/ 

(d) A very reticent approach to vibrato on the violin, not at all to 
our modern taste, is suggested by the following nineteenth-century 

(23 la) Louis Spohr, Violinschule, Vienna, 1832, II, 20: 

'[In the vibrato called oddly by its singer's name of tremolo] the 
deviation from the perfect intonation of the note, should hardly be 
perceptible to the ear. . . . Avoid however its frequent use, or in improper 
places/ [Even in the twentieth century, some hesitation remains as to the 
continuous vibrato.] 

(23 Ib) Carl Flesch, The Art of Violin Playing, transl. F. H. Martens, 
Boston, 1924, ed. of 1939, II, 40: 

'From a purely theoretic standpoint, the vibrato, as a means for securing 
a heightened urge for expression, should only be employed when it is 



musically justifiable [but in practice the continuous vibrato is to be 

A curiously roundabout way of putting it, but one sees what he 
means. In truth a continuous vibrato always is musically justifiable 
provided it is just as continuously adapted to the degree of intensity 
which the music momentarily requires. Totally vibrato-less string tone 
sounds dead in any music. It is just as much an illusion to think that 
early performers preferred it as to think that early singers preferred a 
'white' tone. Exaggerated vibrato makes the tone opaque, and it is 
this opaqueness which is damaging to early music. Sensitive vibrato 
not only can but should be a normal ingredient in performing such 
music: while leaving the tone transparent, it is quite indispensable in 
bringing it to life, as the evidence and practical experience combine in 



B(iii): The Trill 

English: Shake 

Latin: Crispatio 


Italian: Groppo 


French: Cadence 

Pince renverse 



German: Triller 

By a trill is meant a more or less free and rapid alternation of the 
main note with an upper accessory note a tone or semitone above it. 
Trills are as indispensable to baroque music as appoggiaturas, and 
even commoner. Few cadences apart from plagal cadences are complete 
without the conventional trill at least in one of the parts. Cadences are 
an inescapable feature of baroque style, and rather than trying to escape, 
them, it is better to carry them off with conviction, including the almost 
inevitable trill. That means not only starting the trill in standard 
baroque manner with its upper note, but accenting, and often pro- 
longing, that upper note with great assurance and emphasis. Many 
modern performers who are aware of the need to start baroque trills 
with the upper note still do not realise that the entire stress should go 
to it, the remainder of the trill functioning as the merest resolution of 
the strong discord thus introduced. When the trill is taken with its 
proper energy, it relieves the platitudinousness even of the most banal 
of cadences, and adds a glitter of brilliance to those which are already 
interesting in themselves. 


(a) Figures resembling trills occur in the earliest sixteenth-century 
treatises on ornamentation. 



Ex. 84. Sylvestro di Ganassi, Fontegara, Venice, 1535, specimen 
divisions showing the notes of trills: 

Ex. 85. Diego Ortiz, Tratado de glosas, Rome, 1553, specimen 
division showing the notes of a trill: 

r rrn 

(b) Shortly after the middle of the sixteenth century we find several 
specific ornaments taking shape which became prominent in the 
baroque period, among them genuine trills, of which the notation is 
still measured in appearance, but the performance is free and un- 
measured, as the following implies. 

(232) Fray Tomas de Santa Maria, Arte de taner fantasia, Valladolid, 
1565, sect. 8: 

'Take care not to make the redoble too long with the effect of making 
the music clumsy.' 

The redoble is a trill with special preparation; normal trills, under 
the name of quiebro, are also shown, clearly subject to the same 

Ex. 86. Fray Tomas de Santa Maria, Arte de taner fantasia, Valla- 
dolid, 1565, sect. 8, special trill and normal trill: 

RedoHe Quiebro 

Written-out trills apparently measured but meant to be free occur in 
manuscripts and editions not only of the sixteenth but of the seventeenth 
century. The number of repercussions shown is usuaEy but not always 
what the measure requires; it is not always the same in different copies 
of the same passage; some copies may have a sign instead. The intention 
in all cases is to leave the number of repercussions to the discretion of 
the performer: in other words, to indicate not a measured division, but 
a normal trill. 

(233) Girolamo Frescobaldi, Toccatas, Rome, 1614, Preface, sect. 8: 



'You must not divide the trill exactly note for note, but only try to 
make it rapid/ 

(233a) J. A. Herbst, Musica Moderna Prattka, Nuremberg, 1642, 
ed. of 1652/3, p. 59: 

*You beat as many [repercussions] in the trill as you desire [i.e. regardless 
of the notation]/ 

Ex. 87. Giovanni Luca Conforti, Breve et facile maniera d'esserti- 
tarsi afar passaggi, Rome, 1593, written-out trill: 

Groppo di sopra 

Ex. 88. Giulio Caccini, Euridice, 1600, divisions ending (a) with 
written-out trill; (b) with implied trill (probably starting from its 
upper note) : 

Written-out trills occur in the manuscripts of the English Virginalists, 
as well as the double and single strokes, $ and j', of which certainly 
the first and perhaps the second includes the trill (usually a short half- 
trill) among its possible interpretations. It is further a possibility, 
though by no means a certainty, that the double stroke suggests an 
ornament (not necessarily restricted to one kind) made with more 
repercussions than the single stroke suggests. It does not seem to ine 
very probable that either of these signs was meant as a specific indica- 
tion of one ornament only, though this too is by no means impossible. 
The chief arguments against it are first the difficulty of finding two 
separate ornaments to fit equally well all the contexts in which this 
pair of signs are used, and second the inconsistency with which the 
signs appear in the original sources. See App. IV. 

(234) Girolamo Diruta, // Transilvano, II, Venice, 1609, iv, 18: 

'Tremolos [here meaning trills] are to be played at the beginning of a 
Ricercare, or song, or any other composition and again when one hand 



plays several parts and the other only one, you should play tremolos in 
the latter, and then at the organist's convenience and discretion, observing 
always that the tremolo, if played lightly and gracefully and appropriately, 
makes the music live and sound beautiful. 

[p. 19:] 'You should take care to play tremolos with extreme lightness 
and agility [alternating not] with the note below [i.e. as mordents, but] 
with the note above [i.e. as trills], and if you have ever observed players 
of the viol, violin, lute and other string instruments, and even wind 
instruments, you must have noticed that they accompany [the main note] 
of the tremolo with the note above and not with the note below.' 

Ex. 89. Girolamo Diruta, // Transilvano, II, Venice, 1609, acceler- 
ating trill: 

' j r .-. i . 

d- jj. 

The intention in the above example is evidently a gradually acceler- 
ating trill, as it also is in the following. 

Ex. 90. Claudio Monteverdi, Orfeo, 1607, Act III, accelerating 

(c) The trills so far considered, being primarily melodic in their 
function, start indifferently from their lower (main) note or from their 
upper (accessory) note. The latter have slightly more harmonic influ- 
ence, but only by way of coloration. They do not modify the progres- 

(d) The termination which predominates in the trills so far considered 
is a turn, a characteristic and melodically graceful ending which retains 
its association with many trills to the present day. 


(a) The baroque trill proper is a harmonic ornament, and conse- 
quently starts, in all standard cases, from its upper (accessory) note, 
well accented to mark the ensuing modification of the harmony, and 
often to a greater or lesser extent prolonged so as to give this modifi- 
cation still greater prominence. 

A typical late renaissance ornamental resolution of a suspended 
discord might be as follows. 


Ex. 91. Typical late renaissance resolution by way of a trill: 


Played approx 

Here the discord is present in the written progression. But by an 
easy transition, trills became used very early in the baroque period to 
introduce a discord (prepared or otherwise) which is not present in the 
written progression. This discord is the upper accessory of the trill 
itself, which may change a plain dominant 5-3 to tonic 5-3 progression 
into a much more interesting dominant 5-4, through dominant 5-3, to 
tonic 5-3. Here the 4 is the upper note of a trill on the 3. A trill on the 
5 gives dominant 6-3, through dominant 5-3, to tonic 5-3. A double 
trill in thirds on both 5 and 3 gives dominant 6-4, through dominant 
5-3, to tonic 5-3. In each case the resolution is the main note of the 

Ex. 92. Typical baroque trills modifying the written harmony: 

Played approx: 
Written: "^ 


played approx. 



It is because of their modification of the harmony, and particularly 
of cadential harmony, that baroque trills are so often obligatory and 
not merely optional ornaments. 

(235) B. de Bacilly, UArt de Bien Chanter, Paris, 1668, p. 164: 

*[The trill is] one of the most important ornaments, without which the 
melody is very imperfect/ 

We see above that the melodic function was not altogether forgotten; 
and indeed as late as 1676 we find Mace (Mustek's Monument, London, 
p. 103) describing a genuinely lower-note trill for the lute, as well as 
the normal upper-note trill of his period. 

The cadential function, essentially a harmonic function, is, however, 
stressed in the following. 

(236) Tosi, Opinioni, Bologna, 1723, tr. Galliard, London, 1742, p. 

'Whoever has a fine shake, tho' wanting in every other Grace, always 
enjoys the Advantage of conducting himself without giving Distaste to 
the End or Cadence, where for the most part it is very essential/ 

(237) Joachim Quantz, Essay, Berlin, 1752, IX, i: 

'Trills infinitely enhance the rendering of music, give it a great bril- 
liance, and just like appoggiaturas are indispensably necessary to it/ 

(b) The necessity for the cadential trill must be recognised by the 
performer regardless of whether any sign or other hint is present in the 
notation or not; and if a sign is present, regardless of what that sign 
may be, since there are many possible signs misleading or otherwise, 
but there is only one basic species of cadential trill. 

(238) L M. Hotteterre, Prindpes de la flute traversiere, Paris, 1707, 
p. 18: 

'It is necessary to point out that the trills (Cadences ou tremblements) 
are not always marked in musical pieces/ 

A masterly understatement. 


(a) From our modern point of view, all standard baroque trills are 
'prepared' trills, beginning as they do on their upper notes; but in 
baroque terminology, the upper-note start was taken for granted, and 
the trill was called 'unprepared' if this initial upper note was not 
especially prolonged, 'prepared' only if it was very decidedly prolonged. 



Ex. 93. Playford, Introduction, London, 1654 (1660 on), table of 
ornaments, trill: 

u J 

A Backfall 


The above trill is 'unprepared' unless the preceding crotchet G is tied 
to it, in which case it is 'prepared', the crotchet G constituting the 
preparation (a very common method of performance). 

Ex. 94. Purcell (or his ed.), posthumous Lessons, London, 1696, 
unprepared and prepared trills (in the baroque sense) : 

Jean Rousseau (Traite de la Viole, Paris, 1687, p. 76) makes this same 
very usual baroque distinction between trills 'prepared* (avec appuy) 
and 'unprepared' (sans appuy). 

Ex. 95. D'Anglebert, Pieces de clavecin, Paris, 1689: 

fl 4\- .W . 

flr if 1 

TremMcment Tremblement 

^ simple 




(239) J. M. Hotteterre, Principes de la flute tracer siere, Paris, 1707, 

'[Trills begin] on the sound above [and are slurred) without taking 
breath or giving further strokes of the tongue . . . [the preparation may 
extend to] about half the duration of the [main] note, especially in grave 

(b) The rule that the preparation of the trill shall be slurred to it 
(the slur including the whole trill) is a further link between the trill and 
the appoggiatura. Some baroque authorities actually explain the trill as 
a series of reiterated appoggiaturas from above. More important, the 
preparation of the trill is commonly described as an appoggiatura, 
which is, indeed, exactly its musical effect. 

(240) Joachim Quantz, Essay, Berlin, 1752, IX, 7: 

'Each trill starts with the appoggiatura . . . [8:] The appoggiatura is 
often as quick as the other notes which form the trill; for example when 
after a rest there comes a new idea with a trill [i.e. a new entry actually 
starting with a trill]. However this appoggiatura, whether it is long or 

u 177 


short, ought always to be attacked with a stroke of the tongue [on the 
flute]; but the trill and its termination ought to be slurred/ 

(241) C. P. E. Bach, Essay, Berlin, 1753, II, ii, 9: 
['The appoggiatura] is found before cadential trills.' 

(242) Leopold Mozart, Viotinschule, Augsburg, 1756, X, 11: 

If a trill occurs in the middle of a passage [as at Ex. 96 (a)] then not only 
is an appoggiatura made before the trill, but the appoggiatura is held 
throughout half the duration of the note, while the trill is not started till 
the remaining half [as at Ex. (b): t. transcribed as tr throughout]. 

'But if a passage begins with a trill an appoggiatura is hardly heard, and 
is in such a case nothing but a strong attack on the trill [as at Ex. (c)]/ 

Ex. 96. Leopold Mozart, Violinschule, Augsburg, 1756, X, 11: 

The need for preparation is confirmed by other authorities who do 
not use the term appoggiatura for it. The following quotation is 
particularly to be noticed. 

(243) Tosi, Opinioni, Bologna, 1723, tr. Galliard, London, 1742, p. 

'The Shake [Trill], to be beautiful, requires to be prepared [in die 
baroque sense], though on some Occasions, Time or Taste will not 
permit it. But on final Cadences, it [the preparation] is always necessary.' 

(c) The following examples from Couperin are ambiguously notated, 
but can be interpreted, as shown, from the other evidence. 

Ex. 97. Fr. Couperin, Pieces de clavecin, Paris, 1713, table of 
ornaments, trills: 



Trerablement apjpuye et lie 



Trcmblement ouverl 

Interpretation (or in rhy thm as at (c ) ) 




Trembleraent ferae 


Interpretation (or la rfajthm as shown at (6) 

Trenvblement lie sans etre apjpwye 

F_ H (Interpretation.) 


The following example is of unusual interest. The minim G is a long, 
written-out note of preparation (the harmony most probably implied 
being 8-6-4 on a G bass). But the trill itself is assumed to have its 
initial upper note somewhat prolonged (still 8-6-4) before beginning 
its repercussions (on to 7-5-3). The combined result is a preparation 
taking three-quarters of the bar and leaving only a semiquaver for the 
trill and a second (perhaps slightly delayed) semiquaver for its termin- 
ation. It is worth remembering that in the French school, with its taste 
for very long-prepared trills, this might still have been the interpretation 
even if the notation had shown nothing but a semibreve F lasting the 
whole bar through. 

Ex. 98. J.-J. Rousseau, Diet., Paris, 1768, Plate B, trill with 
very long preparation (the second way of writing the interpretation 
is mine, the first is Rousseau's) : 





^ ~ 



L -TT - 

(d) The next example is important because it has been erroneously 
imagined, in modern times, that a baroque trill preceded by a note 
which is the same note as its own accessory note is an exception to the 
normal rule, and should begin on its main note. This is not the case; 



all that happens is that the accessory note begins the trill, as usual, and 
does so either by repeating the previous note, or by being tied to it: 
either interpretation being equally correct and the choice between them 
depending on taste and context. The example concerned is given by 
C. P. E. Bach in illustration of his rule, quoted at (241) above, that 
cadential trills need preparation. 

Ex. 99. C. P. E. Bach, Essay, Berlin, 1753, II, ii, 9, trill prepared 
even when preceded by the same note as its own accessory note (my 
interpretations) : 


4 j jj j ii j. 

3JTS _* g . II 

[interpretation(b) ] 

The correctness of the two alternative interpretations here suggested 
is confirmed by the following, which mentions them both. 
(244) F. W. Marpurg, Anleitung, 2nd ed., Berlin, 1765, 1, ix, 7, p. 55: 

*A trill, wherever it may stand, must begin with its accessory note . . . 
[annotation on p. 56:] If the upper note with which a trill should begin, 
immediately precedes the note to be trilled, it has either to be renewed by 
an ordinary attack, or has, before one starts trilling, to be connected, 
without a new attack, by means of a tie, to the previous note.' 

(e) There are certain ornamental preparations which can be given 
to the trill under suitable circumstances. These are the result of pre- 
fixing another ornament to the trill, and will be discussed in Ch. 
XXVII, 3 below, under the heading of compound ornaments. 

(f ) So general was the baroque view of the initial upper note of the 
trill as in normal circumstances the actual equivalent of an appoggiatura 
that the little note commonly used to indicate an appoggiatura was 
sometimes used as a sign for a (well prepared) trill. Thus at bar 26 of 
J. S. Bach's Cantata 1, Wie schon leuchtet . . . , the second oboe da 
caccia has this appoggiatura sign against tr in the second violin, with 
which it is playing in unison ; and the same thing happens to the first 
oboe da caccia and the first violin later in the bar (passage cited, but 
misunderstood, by Adolf Beyschlag, Die Ornamentik der Musik, 
Leipzig, 1908, ed. of 1953, p. 122). The little quaver C sharp in bar 1 
of the aria 'Erbarme dich' from J. S. Bach's St. Matthew Passion is 
best taken as a prepared trill in this way, the semiquaver A sharp being 
delayed and shortened to a demi-semiquaver in the usual manner. 



(The first ornament is a Slide, for which see Ch. XVI. It Is taken on 
the beat,) 

Ex. 99a. J. S. Bach, appoggiatura sign indicating trill: 


(245) J. M. Hotteterre, Principes de la flute traversiere, Paris, 1707, 
p. 11: 

'The number of repercussions is governed solely by the length of the 

(246) Fr. Couperin, Pieces de Clavecin, Premier Livre, Paris, 1713, 
note on table of ornaments : 

It is the length of the note which should decide the duration of mor- 
dents, of appoggiaturas, and of trills. You must understand by the word 
duration the greater or lesser number of strokes, or repercussions/ 

(247) Fr. Couperin, UArt de toucher, Paris, 1716, ed. of 1717, p. 23: 

4 Although the trills are marked as regular in the table of ornaments in 
my first book, they are nevertheless to begin more slowly than they 

This makes it clear that when Hotteterre and Couperin describe the 
number of repercussions as being governed by the duration of the note, 
they do not mean that an exact number of repercussion is to be 
measured into that duration; they mean that the repercussions are to 
proceed at a suitable unmeasured speed, the actual number being a 
product of speed and length in combination. 

The accelerating trill for which Couperin here asks is confirmed by 
other evidence, and was certainly one of the customary methods of 
performance. It was not, however, by any means the only method, and 
was probably less usual than the following. 

(248) Joachim Quantz, Essay, Berlin, 1752, IX, 5: 

'For trills to be perfectly beautiful, they must be made equal, that is to 
say, of an equal speed and one kept to the same rapidity/ 

As usual, it is the musical context which determines the most suitable 
performance of any given trill. 

(249) Joachim Quantz, Essay, Berlin, 1752, IX, 2: 

'There is no need to make all trills with the same speed. It is necessary 
to adapt yourself not only to the place where you play, but also to the 



piece itself which you have to play. If the place where you play is large, 
and if it reverberates, a rather slow trill will make a better effect than a 
quick trill [and vice versa] ... In addition you must know how to dis- 
tinguish what sort of piece you are playing, so as not to confuse one thing 
with another, which is what happens with a lot of people. In sad pieces 
the trills are made slowly; but in gay pieces they ought to be made more 

[3:] 'With regard to the slowness and quickness, you must not fall into 
any excess. The completely slow trill, only used in French singing, is 
worth as little as the completely quick trill, which the French call bleating 

(250) Leopold Mozart, Violinschule, Augsburg, 1756, X, 7: 

*The slow [trill] is used in sad and slow pieces; the medium in pieces 
which have a lively but yet a moderate and restrained tempo, the rapid in 
pieces which are very lively and full of spirit and movement, and finally 
the accelerating trill is used mostly in cadenzas and with gradual cres- 
cendo .... The trill must above all else not be played too rapidly/ 


(a) Every standard baroque trill (as opposed to half trills) requires 
a termination, which can only be one of two kinds, unless it is compli- 
cated by further ornamentation, for which see Ch. XXV, 3 below. 

(b) The oldest termination was the turned ending to which attention 
has already been drawn in the sixteenth-century examples in 2 above. 
It appears in Playford (Introduction, London, 1654, etc.), in Purcell's 
posthumous Lessons (London, 1696), in Tosi (Opinioni, 1723), in 
Gottlieb Muffat (Componimenti, c. 1736) and many other sources, and 
has in fact never gone out of use. 

(251) Joachim Quantz, Essay, Berlin, 1752, IX, 7: 

'The end of each trill consists of two little notes, which follow the note 
of the trill and which are made at the same speed. . . . Sometimes these 
little notes are written . . . but when there is only the plain note . . . 
both the appoggiatura [preparation] and the termination must be under- 

When the little notes are written, we have to remember that they are 
not necessarily to be taken at their written time value, but as a con- 
tinuation of the trill itself, a point on which Quantz, above, is confirmed 
by other authorities. 

(252) C. P, E. Bach, Essay, Berlin, 1753, II, iii, 13: 

'Trills on notes of a certain length are played with a termination . . . 
[15:] The termination must be played as quickly as the trill itself/ 



This is nearly always the best rule, but it was not universally followed, 
and there are some contents in which one of the occasional variants 
may be preferred. 

Ex. 100. Purcell (or his editor), Lessons, 1696, table of ornaments, 
trill with termination in unusual rhythm : 

Ex. 101. Leopold Mozart, Violinschule, Augsburg, 1756, X, 6, 
trill with termination in most unusual and unattractive rhythm: 

(253) J. C. F. ReUstab, C. P. E. Bach's Anfangsgriinde mil einer 
Anleitung, Berlin, 1790, p. ix: 

'Present fashion has brought it in that the termination shall be slower 
than the trill; and that a little embellishment is added to the termination. 
Formerly this was thought absurd; true connoisseurs may still think so/ 

Ex. 102. Leopold Mozart, Violinschule, Augsburg, 1756, X, 13, 
trill with ornamented turn: 

A*u.. f rJN^i 

Whatever true connoisseurs may have thought, neither this variant 
nor any others replaced the normal termination as standard practice. 
There was, however, one exception which, while cases calling for it do 
not arise very often, was fully in keeping with standard practice when 
they do. 

(254) C. P. E. Bach, Essay, Berlin, 1753, II, iii, 13: 

'Although in rather slow tempos the trills [at Ex. 103] may be given a 
termination (in spite of the fact that the short notes following the dot can 
serve instead) . . . it is not absolutely necessary to insert the termination, 
provided that the dotted notes are trilled for their full duration.' 

Ex. 103. C. P. E. Bach, Essay, Berlin, 1753, II, iii, 13, short notes, 
after trills, which may serve as their terminations: 



The following, according to Quantz, requires a turned ending. 

Ex. 104. Joachim Quantz, Essay, Berlin, 1752, Engl. tr. c. 1790, 
Ex. 21 (a), trill followed by rest but requiring termination: 


(c) The alternative termination, which, for the baroque period itself, 
was as important as the turned ending just described, is a little note of 
anticipation inserted just before the note succeeding to that on which 
the trill is made. 

(255) B. de Bacilly, UArt de Bien Chanter, Paris, 1668, p. 164: 

'The termination ... is a join made between the trill and the note on 
which it is desired to arrive, by means of another note touched very 

[p. 183:] 'Although the composer has not marked on paper the joins 
after the trills ... it is a general rule to assume them, and never to 
suppress them, otherwise the trill will be maimed, and will not be com- 

Where, on the other hand, the composer has written down the note 
of anticipation, it may appear misleadingly, at a greater length than it 
must be given in performance. The normal baroque principles of 
rhythmic alteration (for which see Ch. XLIII below) come into 
operation here, by which a dotted figure may have its dot prolonged 
and its shorter note delayed and shortened, usually with a silence of 
articulation taking part of the value of the prolonged dot. The following 
would be a typical example. 

Ex, 105. Conventional notation implying a trill with termination 
by a short note of anticipation: 

but intended approximately: 

In the above example, the main note of the trill is shown held to a 
length of a dotted quaver before the demisemiquaver silence of articu- 
lation and the demi-semiquaver note of anticipation. This holding of 
the main note, which cannot arise when the trill is terminated by a 
turned ending taken normally at the speed of the trill itself, is the 
meaning of the following rule. 



(256) T. B. s Compleat Muslck- Master, 3rd ed,, London, 1722, ch. Ill: 

'Always let the Proper Note [i.e. the main note of the trill] be dis- 
tinctly hear'd at the last.' 

(257) F. W. Marpurg, Ankitung, 2nd ed., Berlin, 1765, 1, ix, 7, p. 55: 

S A trill . . . must finish, with a certain emphasis, on its main note, for 
this to be properly felt.' 

Ex. 105a. Leopold Mozart, Violinschule, Augsburg, 1756, X, 6, trill 
terminated with note of anticipation : 

Ex. 106 (b) following is the end of a passage which balances that 
ending at Ex. 106 (a) ; they are meant to sound the same (except for any 
flourish added impromptu by the player). The inconsistency in their 
notation is revealing because it shows that a cadential Q may 
normally be taken for J3 , and, of course, given the cadential trill 
which this latter figure invites. 

Ex. 106: G. F. Handel, Deutsche Arien, No. 4, Vn, Obbligato: 



Ex. 107. 'Corelli clash' (a) as written and (b), (c), as intended to 
be performed (approximately) : 




(a) The following pre-baroque half-trills are unprepared in the 
modern sense. 

Ex. 108. Elias Nikolaus Ammerbach, Tabulatur, Leipzig, 1571, 
unprepared half-trills: 



(b) The standard baroque half-trill is prepared in the modern but 
not in the baroque sense of the word : that is to say, it starts with its 
upper note, but this initial upper note is not, as a rule, substantially 
prolonged. There is no termination; the trill ends on its main note, 
held long enough to make a distinct effect (but subject to the total time 
available). The function of the half-trill is scarcely harmonic at all; it 
is to some extent melodic; but it is primarily rhythmic, particularly in 
its shortest examples. 

The least number of repercussions to constitute a trill at all is two, 
giving four notes (one repercussion is an appoggiatura). This double 
repercussion is the half-trill at its purest, in which form its correct 
German name is Pralltriller ('compact trill'). The Pralltriller occurs 
most typically in passages descending by step, as can already be seen 
in the unprepared half-trill at Ex. 108 above. 

(c) The theoretical interpretation of the following Pralltriller, which 
is also the practical interpretation except at too rapid a speed, is as 

Ex. 109. C. P. E. Bach, Essay, Berlin, 1753, II, iii, 30, half-trill 
(Der halbe oder Prall-Triller): 

At a rapid speed, this cannot be executed, and becomes in practice 
the following, which is not theoretically a trill at all, but an upper (i.e. 
inverted) mordent. The German name for this is Schnetter ('snap'); 
and this and the Pralltriller became very confused both in theory and 
in practice during late baroque and post-baroque times. It is quite 
possible that the methods shown at Ex. 120 in 11 below, by which the 
slurring of the ornament to the note before it produces this transforma- 
tion of a Pralltriller (half-trill) into a Schneller (inverted mordent) even 
without the compulsion of a rapid speed, were already coming into 
fashion when C. P. E. Bach wrote in 1753, although he does not himself 
recognise them. But I am not aware of any evidence which would 
entitle us to apply them as far back as the generation of J. S. Bach, 
when no official recognition seems to have been accorded to the 
inverted mordent at all, and the only cases can have been those produced 
by a rapid speed as at Ex. 1 10 next below. 

Ex. 110. Pralltriller becoming Schneller at speed: 

1 Tie omitted in original, but shown correctly in the identical illustration (presumably 
borrowed from C. K E. Bach) by his brother, J. C. ,P. Bach, in the Musikalische Neben- 
stwden, I, Rinteto, 1787. 



(d) The same half-trill may be taken without the tie; but it is just as 
subject to the effect of great speed. 

Ex. 111. (a) half-trill not tied, (b) as performed at moderate speed, 
(c) and (d) as transformed into an upper mordent by increasing 

The German terms Pralltriller for (b) and Schnelkr for (c) or (d) are 
equally applicable to these untied versions, where they have, however, 
been equally subject to confusion. 

(e) Whether tied or not, the short half-trill {Pralltriller) is always 
performed (even in slow movements) with the greatest attainable 

(258) C. P. E. Bach, Essay, Berlin, 1753, II, iii, 32: 

c The half-trill or Pralltriller (der halbe oder Prall-Triller) . . . must 
literally crackle . . . with such extreme speed that the separate notes will 
only be heard with difficulty . . . 

[34:] "The half trill or Pralltriller only occurs on a descending step of a 
second regardless of whether the interval is produced by an appoggiatura 
or by large [i.e. main or written] notes . . . when it occurs over a note 
prolonged by a pause, the appoggiatura is held rather long and the trill is 
smartly snapped as the fingers draw away from the keys/ 

Where the written notes are short, the following example shows with 
what extreme velocity the ornament has to be carried off, and how 
inevitably it becomes changed into the upper (inverted) mordent or 
Schnelkr as soon as the tempo becomes rapid. 

Ex. 112. C. P. E. Bach, Essay, Berlin, 1753, II, iii, 35, (a) half-trill, 
(b) changed into an inverted mordent by the effect of speed, (c) as 
performed if the speed permits (my interpretations): 

Of the following examples, the first should be taken as normal half- 
trills (Pralltriller), but the second, not only on account of the speed but 
also from the presence of the slur, can in practice only be satisfactorily 
executed as inverted mordents (Schnelkr). (Leopold Mozart states in a 
footnote to the edition of 1787 that 'all these trills are without turns'.) 



Ex. 113. Leopold Mozart, Violmschule, Augsburg, 1756, X, 16-17, 
(a) half-trills and (b) inverted mordents (my interpretations). 

(f) But while it is both necessary and correct to allow the fastest 
half-trills to be transformed into inverted mordents by the effects of 
speed, it is neither correct nor musically desirable under any circum- 
stances to allow them to anticipate the beat. 

(g) Half-trills can also be used in a more leisurely manner than the 
Berlin school of C. P. E. Bach preferred. They can, moreover, on 
sufficiently long notes, be allowed more repercussions, provided time is 
still left to dwell on the last appearance of the main note sufficiently 
to prevent their turning into hybrid, unterminated full trills. The 
following is a typical half-trill of greater length. 

Ex. 114. Long half-trill: 


(a) A special case arises when trills are found, or are placed by the 
performer, over a number of notes in succession. We may call them 
continuous trills. 

(b) The continuous trill may be of any length from very short to 
very long; but whatever its length, it normally fills out the entire 
duration of the note on which it is placed. We therefore cannot regard 
it, even in its shortest examples, as a half-trill. It is a full trill of which 
the behaviour is in some respects irregular. 

(c) The purpose of the continuous trill is essentially to add brilliance. 
Its shortest examples serve primarily a rhythmic function; its longer 
examples might be called melodic, although their primary effect is one 
of coloration. In no case does the continuous trill serve a primarily 
harmonic function. It starts as usual with the upper note, and this is 
well accented in the shortest examples, but much less accented in the 
longer examples. And however long the trill, the initial upper note is 
not prolonged, but is taken at the same speed as the remaining 

(d) In spite of the fact that the continuous trill fills out the entire 
duration of the note, there is no necessity for a termination, though 



under some circumstances a turned ending may be supplied at the 
performer's option. 
(259) Leopold Mozart, Violinschule, Augsburg, 1756, X, 22: 

'These rising and falling [continuous] trills [may be fingered I, which 
makes them without a turn: Ex. 115 (a) and (b)]. 

[25:] 'With notes [not taken stepwise but] lying some distance apart, 
continuous trills may be used, but this is seldom feasible in a quick allegro, 
and then normally only in cadenzas. [Ex. 115 (c)]. 

[Added in ed. of 1787:] 'These continuous trills used on notes lying 
some distance apart are best taken with turned ending. Also the rising and 
falling trills shown [at Ex. 115 (a), (b) and (c)] can be taken with turns if 
the tempo is very slow . . . but the turn must be rapid and fierce. 

[26 :] 'There is a species of rising and falling trill each of whose [main] 
notes [has instead of the turn] a quick leap to the open string below. 
[Ex. 115 (d)]. In such cases the trill should be held on [virtually] as long as 
if there were only one note, and the leap down should be delayed so as 
to be scarcely heard/ 

Ex. 115. Leopold Mozart, Violinschule, Augsburg, 1756, X, 22ff. 9 
continuous trills: 

. A ttj 




[also unslurred] 



{The first G$ is presumably 
& misprint ibr EJ 

(e) The longest continuous trills are those found primarily in harpsi- 
chord music on notes so long that they would die away unduly if the 
sound were not kept in being in this way. Similar trills are also found 
for wind and string instruments, either to match a harpsichord part 
which includes them, or purely as an effect of colour. They do, more- 
over, considerably intensify the melodic line, wherever they are found. 

(260) C. P. E. Bach, Essay, Berlin, 1753, 1, ii, 5: 

'The sign [a wavy line] is extended when it appears over long notes. 
[The trill] always begins on the note above the main note . . . [6 :] Some- 
times two little notes from below [i.e. the turned ending] are added. . . .' 



A series of very long trills (perhaps each a full bar in length) may 
either be given turned endings (at the same speed as the trill, and 
therefore not very conspicuous), or they may be run straight into one 
another. But if the last trill ends with the bar (or half-bar) it is generally 
better to give at least this trill a turned ending before passing on to the 
continuation of the melodic line. If, on the other hand, the trill is tied 
over to the first beat of the next bar, it is generally the most effective 
solution to dispense with a termination, ending the trill with this beat, 
and holding the beat itself as a plain note: in other words, to treat the 
trill as a greatly prolonged half-trill. 

It is also possible to treat all the trills in the series as prolonged half- 
trills, dispensing with a termination, and holding the main note plain 
for perhaps the last third or quarter of its length. 


The ribattuta (Ger. Zuriickschlag) is an ornament of the Caccini- 
Monteverdi school which remained in some use at least to the end of 
the baroque period. It is in effect a slowly accelerating trill of which the 
rhythm is uneven until the repercussions reach a certain speed, by 
which time they have gradually become even and take the form of an 
ordinary trill either in the old Italian sense of tremolo, or in the later 
sense in which we use the word trill. Originally a singer's ornament, but 
subsequently borrowed by instrumentalists, the ribattuta always served 
a melodic function, and it begins on its lower note. It is really effective 
only for the voice; yet C. P. E. Bach (Essay, Berlin, 1753, II, iii, 25) 
shows it for the harpsichord, and Leopold Mozart (Violinschule, Augs- 
burg, 1756, X, 5) for the violin. Mattheson gives it only for the voice, 
and says (in 1739) that he has not found it in the published treatises. 

Ex. 116. Johann Mattheson, Vollkommene Capellmeister, Ham- 
burg, 1739, Pt. II, Ch. 3, 48, ribattuta, leading to trill (the whole 
described as a species of 'tenuta') : 

[performed as a gradual acceleration] 


(a) The typical and intrinsically powerful baroque cadence by 
dominant discord and resolution passed at last into transition and 
eventual decline in the subsequent period; and with it the operative 
purpose of the harmonic trill. As early as Daniel Gottlob Turk (Kla- 
vierschule, 1789) we find a hint that exceptions to the upper-note start 
occurred, though he does not show any such exceptions in his examples. 



J. A. Hiiler (Anweisung zum Violinspielen, 1792) neither mentions nor 

shows any exception to the baroque rule; J.-B. Cartier makes one 
exception, and for a very interesting reason : 
(261) J.-B. Cartier, UArt du Violon, Paris [1798], ch. on ornaments: 

*It is necessary to use it often, for if we made continual mordents and 
trills, without ever perceiving the main note, the melody would be too 

Ex. 117. J.-B. Cartier, UArt du Violon, Paris [1798], delayed trill: 

Tcnuc sur la iSot 

This somewhat illogical example appears to give first a form of main- 
note start (but when the trill really arrives it does so with an upper-note 
start), and second a normal upper-note start as an alternative rendering, 

Louis Adam (Methode, 1798) gives a lower-note start only on a long 
cadence of special type, dementi (Introduction, c. 1803) makes several 
exceptions, but otherwise respects the baroque rule. Pollini (Metodo 
used at Milan Conservatory, 2nd ed. 1811); Cramer (84 Studies, c. 
1810); Rode, Kreutzer (to whom Beethoven dedicated the Kreutzer 
sonata) and Baillot with their famous Methode de violon (drawn up for 
the Paris Conservatoire, chiefly by Baillot, whose own Art du violon 
was published in 1834): all these retained the upper-note start without 

But on the other hand, Hummel (Anweisung, 1828), Spohr (Violin- 
schule, 1832) and Czerny (Pianoforte School, 1839), argued in favour 
of a lower-note start to all trills, on the same reasonable grounds 
already used by Cartier at (261) above to justify his single case: i.e. 
that for melodic purposes the main note ought to be more strongly 
impressed on the ear than the accessory note. 

The trill was now again, in short, what it had previously been in the 
sixteenth century: a primarily melodic ornament. As such, it should 
perhaps more logically have a main-note start; but this is not a pressing 
necessity, as the upper-note start is for a harmonic trill. Even in the 
second half of the nineteenth century we can still find evidence for the 
upper-note start as an equally recognised alternative to the main-note 

(262) F. de Courcy, Art of Singing, London, c. 1868, p. 95: 

'[In trills, the main] note itself, or the [accessory] note above, may be 
begun with and given the accent to. . . .' 

In must be clearly understood that the upper-note start remained 
standard at least until the second quarter of the nineteenth century. 



Hummel, for example, was proposing to standardise the main-note 
start as an innovation in 1828, a year after Beethoven's death. The 
interpretation, then, which Haydn and Mozart, Beethoven and Schubert 
intended for their trills, and which their contemporary performers 
gave them, was with an upper-note start. Our present interpretation 
with a main-note start is contrary to the composer's intention. 

On the other hand, while the initial upper-note will always take the 
accent normal to its position in the bar, and may frequently be given 
added accentuation, it will not ordinarily be prolonged. Such prolonga- 
tion is only associated with trills used primarily for harmonic purposes; 
and that was only exceptionally the case in nineteenth century music. 

In the adagio of Beethoven's quartet Op. 127 (bar 55, Vn. II) a single 
appoggiatura appears which may be taken as confirming a normal 
initial upper note to the trill ; but the same treatment must be given to 
the many corresponding trills which are not thus marked. The following 
example supports such a treatment. 

Ex. 118. I. Pleyel, Methode de pfte., early 19th cent., trill with 
prolonged initial upper note shown as appoggiatura: 




(b) The body of the trill remained what it still is: the same as in the 
baroque period. 

(c) The termination of standard trills of the late-eighteenth century 
and the nineteenth century was the turned ending. There seems to be 
a rare unanimity among the authorities on this point, including all 
those so far cited in the present section. It will be sufficient to quote 
three passages, the first two because they are so plainly put, the third 
for its late date. 

(263) Turk, Klavierschule, Leipzig and Halle, 1789, ed. of 1802, p. 27: 

'The Turn is made to the Shake trill even without a Mark, when the 
length of the Note admits of it.' 

(264) Hummel, Anweisung, Vienna, 1828, III, p. 387: 

'Every trill must end with a turned ending, whether this is marked or 
not. . . . Except for rare and special effect it is at the same speed as the 

(265) F. de Courcy, Art of Singing, London, c. 1868, p. 95: 
'Shakes [trills] generally conclude with a turn. . . .' 

The alternative ending by a little note of anticipation was still 
described by the notoriously conservative J.-B. Cartier (UArt du 



Violon, Paris [1798]) as "how one ordinarily and most naturally finishes' 
a trill; but the balance of evidence suggests that this was now the 
exception, and the turned ending the rule. Hummel (Anwefsung, 1828, 
III, 387) certainly treats it so. 

It thus appears that all the hundreds upon hundreds of standard 
trills in Haydn, Mozart, Schubert or Beethoven are meant to be taken 
with a turned ending (performed at the same speed as the rest of the 
trill) in spite of the fact that some are so marked while others are not. 
A typical instance occurs in the opening phrase of Beethoven's violin 
sonata Op. 96. His fourth violin sonata, last movement, second subject 
shows a trill marked with a turn in the recapitulation, but without at 
its first appearance; they should both be the same, Le. with the turn. 
The scherzo of his posthumous quartet Op. 127 has numerous trills of 
which some but not others are shown with turned endings; but they 
all require turned endings, while some additional trills not marked at 
all may be supplied to complete the pattern (e.g. Vn. II, bar 10, cf. bar 
281; Viola, bar 20, cf. bar 291). Later in the nineteenth century, it 
remains by far the strongest probability that trills shown without any 
marked termination should be given a turned ending by the performer. 
The unterminated trill appears to be an entirely modern innovation. 
There is no such thing as an unterminated standard trill either in 
baroque or in classical music. 

(d) The half-trill, on the other hand, continued to be unterminated; 
it also continued to start on its upper-note (ordinarily unprolonged), 
with much the same exceptions due to speed. 

The distinction between the half-trill and the full trill could not be 
more clearly shown than in the following examples. 

Ex. 119. Busby, Complete Dictionary of Music, London, [1786], 
s.v. 'Shake', (a) half-trill, (b) full trill: 

Passing Shake. Turned Shake. 

C. P. E. Bach's rather subtle tied PralltriUer (Ex. 109 above), besides 
being impossible at a rapid speed, is by no means easy to keep intact 
even at a slower speed; and his treatment tended to be replaced by the 

Ex. 120. L'Abbe le fils, Principes du Violon, Paris, c. 1761, (a) 
Schneller, (b) PralltriUer: 

(a) as but (1)) as 

ffi r f f iffi. iiJJj|J^j JH 




In both cases the ornament falls on the beat, as usual. But at (a), 
the effect of the slur is to make the ornament sound as if it already had 
an upper-note preparation in the first note of the bar, and did not 
need another. Its transformation into what is actually a slurred 
Schneller (inverted mordent) has therefore a natural and satisfactory 
effect. At (b), on the other hand, the first note of the bar, not being 
slurred to the ornament, does not sound like a preparation for it, and 
it is therefore natural to supply one in the ordinary way. This keeps the 
Pralltriller (half-trill) intact. 

D. G. Turk (Klavierschule, 1789) still shows C. P. E. Bach's tied 
Pralltriller as the correct interpretation of the sign ~~ in adagio; for 
andante he shows the Schneller^ but includes it under Pralltriller instead 
of giving it its own proper name. Subsequent authorities confused both 
these and other names, as well as a variety of signs, to a quite remark- 
able extent, and it is not surprising that this confusion has persisted, 
and somewhat thickened, in modern times. We have no such excuse for 
taking these ornaments before the beat, or for beginning them before 
the beat, since all the authorities appear to assume that they fall on the 
beat, and a number of them including J. A. Hiller (Anweisung zum 
Violinspielen, 1792), J. H. Knecht (Methodenbuch, early 19th century) 
and I. Pleyel (Methode depfte., early 19th cent.) actually state the fact 
or make it clear from their examples. 

The inverted mordent in its own right (and not merely as a Pralltriller 
transformed into a Schneller) will receive full attention in Ch. XXI, 
3 below. 

The half-trill in its special aspect as a Pralltriller may be treated 
either as at Exx. 109 and 11 1 in 8 above, or as at Ex. 120 above in the 
present section: this remains true for any music in which it makes a 
genuine appearance. 

The half-trill in its more genuine aspect as a trill of two or more 
repercussions but ending, without termination, on its own main note 
held long enough to be distinctly heard, should normally be treated in 
classical music in the same way as in baroque music. It is not unknown 
with a main-note start (e.g. in J. H. Knecht, Methodenbuch, early 19th 
century), but in this form it loses much of its true character as a curtailed 
trill, and becomes in effect a more or less prolonged inverted mordent. 

(e) The continuous trill remained as in baroque times in so far as 
it was still used by virtuoso singers, violinists and others. 

(f) The ribattuta was still described by J. D. Andersch (Musikalisches 
Woerterbuch, 1829), on baroque lines. 



B (iv): The Mordent 

English: Beat 


Open Shake 

Latin: Tremulus 

Italian: Mordente 
French: Battement 





German: Beisser 

Mordant, Mordent 


By a mordent is meant a more or less free and rapid alternation of the 
main note with a lower accessory note a tone or semitone below it. 

The name means 'biting' and a certain ferocity is normally associated 
with the shorter mordents, of which the function is essentially rhythmic, 
while the longer mordents, which have some melodic influence, tend 
to be a little smoother. The ornament is extremely valuable throughout 
baroque music. 


(a) A single mordent has only one repercussion, and is the sharpest 
of rhythmic ornaments except for the acciaccatura. 

(b) A double mordent has two repercussions, and is still primarily 
a rhythmic ornament. 

(c) A continued mordent may have any number of repercussions, 
lasting up to several bars, and like the longest trills, sustains and 
intensifies the melodic line and adds colour to the texture. 

Ex. 121. Fr. Couperin, UArt de toucher le clavecin, Paris, 1716, 
ed. of 1717, (a) single, (b) double, (c) continued mordents: 




-Q \~-~- ===. 

^-TT " : - } 


ft performed 


(a) A single mordent may occur inverted: i.e. alternating with an 
upper instead of with a lower accessory note. 

These are the notes of a trill rather than of a mordent; but being so 
short, the ornament sounds like a mordent and not like a trill. 

The correct German name for this is Schneller; the French name is 
pined renverse or reverse (reversed mordent). We have seen in Ch. XX, 
8 and 11 (d) above, that if circumstances rob a Pralltriller of its initial 
note, it becomes a Schneller. Nevertheless the Pralltriller is not an in- 
verted mordent; it is long enough to sound like a prepared half-trill, 
and that is what it is. The two ornaments are distinct, but since they 
only differ by one brief note and the Schneller may under certain 
circumstances replace the Pralltriller in practice, it is not surprising that 
they became, as they still are, seriously confused. 

(b) There are in practice no other inverted mordents, since to invert 
a double or a continued mordent is simply to turn it into a half-trill or 
a trill 


(a) Mordents may be used in many positions, but the following 
pointers will be found valuable. 

(266) C. P. E. Bach, Essay, Berlin, 1753, II, v, 4: 

'The mordent is particularly effective in an ascent by step or by leap. 
It appears seldom in a descent by leap, and never on descending steps of a 

[14:] 'The mordent is the opposite of the half-trill. This latter may be 
used on a step only in descent, which is just where a mordent is unsuitable/ 

This rule (i.e. that a mordent is effective on a step only in ascent) is 
reliable for pre-baroque as well as for baroque music, and if the reader 
will make the experiment of trying it in reverse, he will see why; for 
the ornament thus used weakly anticipates the following note. An 
inverted mordent, however, behaves like a half-trill in this respect (i.e. 
it is effective on a step only in descent). 

(b) The following are more hints than rules, but as such are widely 


(267) C. R E. Bach, Essay* Berlin, 1753, II, v, 9: 

'Mordents, especially shore ones, lend brilliance to leaping, detached 

[10:] 'The mordent Is of all ornaments the most freely Introduced by 
the performer into the bass, especially on notes at the highest point of a 
phrase, reached either by step or by leap, at cadences and other places, 
especially when the next note stands at an octave below/ 


Ex. 122. Fray Tomas de Santa Maria, Arte de taner fantasia, 
Valladolid, 1565, sect. 8, (a) inverted (upper) and (b) standard 
(lower) mordents ('single quiebro'): 

A (a) (M 

Ex. 123. E. N. Ammerbach, Tabulator, Leipzig, 1571, double 


(a) The inverted mordent, which appears on equal terms with the 
standard mordent in 5, Ex. 122 above, disappears from baroque music, 
and does not reappear until the emergence of the Schneller in C. P. E. 
Bach and other authorities of the mid-eighteenth century, for which 
see Ch. XX, 8 and 11 (d) above. 

(268) F. W. Marpurg (ed.), Kritische Brief e, Berlin, 30 May, 1759, 
letter 2: 

'Ask your honest Mr. Amisallos whether anyone knew anything of the 
Pralltriller or Schneller in his youth/ 

This means that the not uncommon modern practice of relying on 
the inverted (upper) mordent as the chief ornament for baroque music 
is incorrect. The standard baroque mordent is the lower mordent. 

(b) Mordents are among the probable renderings for the English 
Virginalists' at the turn of the sixteenth and seventeenth centuries, 
and perhaps also for / (see Ch. XX, 2 for the possibility that the double 
stroke may suggest more repercussions than the single stroke; but the 
most useful mordent in this music is the single mordent, just as the 
most useful trill, except at cadences, is the half-trill). 

(c) Later in the baroque period, mordents are among the ornaments 
suggested by the ubiquitous + and x . 



(d) At any stage In the baroque period, mordents are among the 
ornaments most freely at the disposal of the performer when not 
marked. This applies as much to singers as to instrumentalists ; but see 
(277) in 9 below. 


(a) A mordent is diatonic when its accessory (lower) note is taken 
at the interval of a tone or of a semitone from its main (upper) note in 
accordance with the key in which the passage stands. 

If this key is not the main key of the piece, as shown in the key 
signature, an accidental may have to be supplied, which may or may 
not be shown in writing: most usually not. If it is not shown, the 
performer must make the implied modification. 

(b) A mordent is chromatic when its accessory (lower) note is taken 
at the interval of a semitone from its main (upper) note although the 
key of the passage would make this interval a tone. 

Here, too, the requisite accidental may or may not be shown, but is 
most usually not. If it is not shown, the choice between a diatonic and 
a chromatic mordent rests with the performer. 

(c) Both diatonic and chromatic mordents were freely used in 
baroque music, though different authorities state somewhat differing 

(269) Thomas Mace, Mustek's Monument, London, 1676, p. 105: 

'Continued mordent always into a Half-Note beneath. 9 

(270) Jean Rousseau, Traitede la Viole, Paris, 1687, p. 88: 

'The Mordent is ordinarily [but not invariably] made on the second 
Note of an ascending Semitone. 5 

(271) GeorgMuffat, Apparatus musico-organisticus, 1690, preface: 

'A mordent [is made with] the lower accessory note, which is often (if 
the ear does not forbid it) a whole tone below/ 

(272) T. B. ? Compleat Mustek-Master, 3rd ed., London, 1722, Ch. II: 

*A Beat [mordent] ... is always from the half Note below. 
'[But Ch. Ill:] A Beat . . . proceeds from the Note, or half Note next 

Ex. 124. Gottlieb (TheopM) MufFat, Componimenti, Augsburg, 
71735, semitone mordents, diatonic and chromatic: 




Ex. 125. Johann Mattheson, Vollkommene Capellmeister, Ham- 
burg, 1739, Pt. II, Ch. 3, 56, chromatic mordent: 


So if i? written 

'.So, more or IMS, it is oiii* 

The notation above is not mathematically logical, but the intention 
is clear. The semitone is here effected by an accidental sharp; but 
though not in the key-signature, it is diatonic to the key of G in which 
the example evidently stands. 

(273) Johann Mattheson, Vollkommene Capellmeister, Hamburg, 

'The lower tone or semitone (according to the key) must be touched/ 

(274) C. P. E. Bach, Essay, Berlin, 1753, H, v, 11 : 

'With regard to accidentals this ornament adapts itself to its context in 
the same way as the trill. Its brilliance is frequently enhanced by raising 
its lower [i.e. accessory] note, as [at Ex. 126]. * 

Ex. 126. C. P. E. Bach, Essay, Berlin, 1753, II, v, 11, chromatic 

C. P. E. Bach's fingering here shows that the notes of all three 
mordents above are the same: i.e. single, standard (lower) mordents. 
The sharp, though it stands above the sign, indicates D sharp. In the 
ordinary course of events we should not expect to see the accidental 
marked at all even if the mordent was. The performer is intended to 
use his own judgment. 

(274a) Leopold Mozart, Violinschule, Augsburg, 1756, XI, 16: 

'The continued mordent [is] always from the semitone/ 
Ex. 127. L. Mozart, Violinschule, Augsburg, 1756, XI, 9, 16: 

(a) unprepared single mordent; (b) unprepared double mordent; 
(c) prepared continued mordent: 



(a) The normal situation with regard to the standard mordent of 
any period (and the inverted mordent at such periods as have used it) 
is that it starts unprepared, on its main (upper) note, which thus takes 
the accent. This accent may often be considerably stressed, particularly 
with the shorter mordents; but the note itself is not prolonged. 

(b) Occasionally, however, a mordent may be prepared by its 
accessory (lower) note, which then receives the accent, but without 
being prolonged: see Ex. 127 (c) above. 

(c) When, on the other hand, this initial lower note is not only 
accented but also prolonged, it was treated by the baroque authorities 
not as the preparation to the mordent, but as an appoggiatura from 
below with a mordent added to its resolution just as a trill with its 
initial upper note prolonged was sometimes treated as an appoggiatura 
from above with a trill added to its resolution. But whereas preparation 
(whether prolonged or otherwise) is an integral part of the baroque 
trill, it has never been an integral part of the mordent. For this reason, 
the combination of appogiatura and mordent will be treated here on 
the baroque precedent as a compound ornament, for which see Ch. 
XXV, 2. 


A long continued mordent is not a rhythmic ornament, but an inten- 
sification and coloration of the texture. Like the long trill which it so 
much resembles, it may be taken at a very steady rate of repercussion. 
But the shorter mordents, and above all the single mordent, are 
rhythmic ornaments, and their repercussions are normally taken at a 
very rapid speed. 

(275) Le Sieur de Machy, Pieces de violle, Paris, 1685: 

6 [To make a mordent,] lift the finger from the [main] note as soon as 
sounded, and replace it at the same time.' 

(276) Jean Rousseau, Traite de la viole, Paris, 1687, p. 87: 

'[To make a mordent,] the finger stopping a Note first makes two or 
three little repercussions more brilliantly and rapidly than a trill, and then 
remains on the Fret/ 

(277) Johann Mattheson, Vollkommene Capellmeister, Hamburg, 
1739, Pt. II, Ch. 3, 55: 

'In playing, [mordents] can be variously used, in singing only in one 
way [i.e. as a single mordent]. First, the written main note is sounded; 
then the tone or semitone below (according to the key) must be touched 



upon, and left for the main note with such speed that these three notes 
make one sound which, as it were, hesitates a little, is delayed by some- 
thing, gently collides with something. From which it should be obvious 
that the Mordent has not to divide or break anything, but rather joins and 
unites the sounds.' 


There is no termination to any normal mordent. The ornament ends 
on its own main note, out of which the single mordent takes very little 
time, and the double mordent not much more, but the continued 
mordent the greater part. 

(278) Couperin, UArt de toucher, Paris, 1 7 1 6, ed. of 1 7 1 7, 4 Agremens* : 

'Every mordent must end on the [main] note on which it stands . . . 
the repercussions and the note at the end must all be included within the 
[total] time-value of the main note.' 

(279) C. P. E. Bach, Essay, Berlin, 1753, II, v, 8: 

'When mordents are continued to fill out a note, a small portion of its 
original duration must still be left plain, since even the best used mordent 
sounds wretched when carried like a trill straight into the next note/ 


(a) Apart from confusing the Pralltriller and the Schneller in a 
manner subsequently very common, Turk (Klavierschuk, 1789) gives a 
clear selection of normal mordents on the baroque pattern. He includes 
prepared mordents, but under the separate name battements. 

Ex. 128. D. G. Turk, Klavierschule, 1789, prepared mordents: 



dementi (Introduction to the Pianoforte, 1801) contrasts with his 
normal trills (upper-note start) a 'short shake beginning by the note 
itself . . . transient or passing shakes', which are in fact inverted 
mordents; but he also gives under the name of 'beat' normal lower 
mordents both prepared and unprepared. On an ascending step he 
recommends taking the accessory (lower) note of the mordent at the 
same pitch (usually diatonic) as the preceding note; but on leaps or 
first notes of passage, he recommends taking a semitone interval 
whether diatonically or chromatically. Hummel (Anweisung zum Piano- 
fortespiel, 1828), Spohr (Violinschule, 1832) and Czerny (Pianoforte 



School, 1839) all give inverted (upper) mordents under a variety of 
names and signs. From the second quarter of the nineteenth century 
onwards, the normal assumption appears to be that an inverted (upper) 
mordent and not a standard (lower) mordent is intended. 

(b) Passages occur in the nineteenth century romantics where an 
upper mordent appears written out in little notes before a main note 
bearing a sign of accentuation, or immediately before a bar-line. On 
the assumption that this notation is authentic, the intention is pre- 
sumably to take the mordent before the beat. But this is not natural to 
the ornament; no question of it can arise in classical any more than in 
baroque music. 



C (i): Passing Notes and 
C (ii): Changing Notes 



The ornaments here grouped together under the loose heading of 
passing notes are those which connect two disjunct main notes by a 
more or less conjunct movement. 


We saw in Ch. XVIII that in France and Germany at the middle of 
the eighteenth century a temporary fashion arose for slurring an other- 
wise ordinary passing note to the following main note instead of to the 
previous one; and that the authorities recommending it treated it under 
the rather unconvincing heading of Passing Appoggiatura. 

When a similar passing note is slurred, as it is much more natural 
for it to be slurred, to the previous note, it is and behaves in every way 
like an ordinary passing note. This ornament is so normal that it does 
not even possess a specific name, being merely one of the most obvious 
of the figures by which disjunct notes a third apart are connected in 
free ornamentation. It became in effect, however, if not in name, a 
specific ornament, and one very commonly introduced by the performer 
without indication in the written text. 

Ex. 129. Passing notes introduced as ornaments: 

written might be performed 


English: Run 
Italian: Tirata 
French: Cascata 



The above are names more or less specifically applied to a series of 
passing notes connecting two disjunct notes more than a third apart. 
When the series is itself conjunct, these are strictly passing notes; when 
it is partially disjunct, the passing notes are called irregular or free. 

Ex. 130. Johann Mattheson, VottkommeneCapettmeister, Hamburg, 
1739, Pt. II, Ch. 3,44, tirata: 

written nitwit be performed 

v ^-^ 4 ' ua - 1 - T1 ^ 1 - * ^4j iiiU; g J u 

A few further examples are given below; but it is obvious that we 
are here on the very borders of free ornamentation. 

Ex. 131. Leopold Mozart, Violinschule, Augsburg, 1756, XI, 20-21, 
tiratas, (a), (b), (d) strict, (c) free, (e) strict but chromatic: 


*. . Molto allegro 
^ ? f& i 9" 




The ornaments here grouped together under the equally loose heading 
of changing notes are those which move in a variety of patterns around 
or between main notes which may be the same, or conjunct, or disjunct. 


An ordinary note of anticipation, slurred to the previous note, and 
taken very lightly, was commonly used and sometimes described as a 
specific ornament (Old English cadent). 



Ex. 132. Playford, Introduction, London, 1654 (1660 on), note of 
anticipation used as an ornament; 

^_^ [i. e. often approx.J 

Ex. 133. F. W. Marpurg, Principes du Clavecin, 1756, Tab. IV, 
notes of anticipation used as ornaments: 


English: Acute 


Springer, Spinger 
French: Accent 


German: Nachschlag 

This ornament, the simplest of true changing notes, is among the 
various accenti given by ML Praetorius (Syntagma, III, 1619), and re- 
mained popular throughout the baroque period. It Is included by Play- 
ford (Introduction, 1654), Simpson (Division-Violist, 1659), Mace 
(Mustek's Monument, 1676), Jean Rousseau (Traite de la mole, 1687), 
Georg Muffat (Florilegium I, 1695), Couperin (Pieces, 1713), Marpurg 
(Principes du clavecin, 1756), Jean- Jacques Rousseau (Dictionnaire, 
1768) and others. 

Ex. 134. Playford, Introduction, 1654 (1660 on), springer: 


A Springer 

Ex. 135. F. W. Marpurg, Principes, 1756, Tab. IV, springers: 

-S 1 


LJ r ii u r 


This name was applied to an even wider variety of ornaments than 
most, including some forms of trills and turns : it means a 'cluster' of 



notes. By the end of the baroque period, however, it had acquired a 
slightly more definite meaning of which the following is representative. 

Ex. 136. Leopold Mozart, Violinschule, 1756, XI, IS.groppi: 


These never very distinct groups of ornaments (to which others 
equally ill-defined might be added) soon lost what specific status they 
possessed, and in post-baroque music appear mostly as ordinary 
written-out figuration as indeed they often do in baroque music also. 



C(iii): The Turn 

English: Single Relish 

Italian: Circolo mezzo 


French: Brisee 




Tour de gosier 
German: DoppelscMag 



(a) The turn is an alternation of the main note with both an upper and 
a lower subsidiary. It is thus a group of changing notes, or groppo (see 
Ch. XXII, 7), which acquired independence as a specific ornament. 

(b) The turn, a most beautiful ornament, has always been subject to 
variations of form and rhythm; but it has not significantly varied 
between different periods. 

By the middle of the nineteenth century it began to decline as a 
specific ornament, but remained just as important as an ingredient in 
the written-out melodic line. Wagner used the sign for it in early life, 
but later wrote it out (as in Brynhilde's main theme in the Ring). 
Mahler made it almost as much a feature of his idiom as the long 
appoggiatuxa (see the last movement of his ninth symphony). 

(c) The main distinction is between accented turns and unaccented 
turns. The former take the beat; the latter fall between beats. The 
difference is sometimes shown by the exact vertical placing of the sign, 
but this cannot be relied upon; and, of course, the ornament may be 
freely used where no sign exists to give any such assistance. 

(d) The upper or standard turn begins on its upper auxiliary, passes 
through its main note, touches its lower auxiliary, and returns to its 
main note. 

The lower or inverted turn begins on its lower auxiliary, passes 
through its main note, touches its upper auxiliary, and returns to its 
main note. 

There is also a main-note or five-note turn, which begins on its main 



note, touches its upper auxiliary, passes through its main note, touches 
its lower auxiliary, and returns to its main note. 

The theoretically possible inverted main-note turn does not seem to 

Upper, lower and main-note turns can all of them occur either in 
accented or in unaccented form. 


Ex. 137. Diego Ortiz, Tratado de glosas, Rome, 1553, accented 
upper turn: 

* Plain, notes Division [>,,} 

r T r 

Ex. 138. Etienne Loulie, Principes de musique, Paris, 1696, Amster- 
dam ed. of 1698, accented upper turn: 

Ex. 139. Fr. Couperin, Pieces de clavecin, I, Paris, 1713, accented 
upper turn : 



Ex. 140. J. S. Bach, Clavier-Buchlein, 1720, accented upper turn: 

Ex. 141. C. P. E. Bach, Essay, Berlin, 1753, II, iv, 1, 34 and 36, 
(a) accented upper turns and (b), (c), accented main-note turns : 



Ex. 142. F. W. Marpurg, Principes, Berlin, 1756, (a) accented 
upper turns, (b) accented lower turns, (c) tied upper turn, (d) accented 
main-note turns: 



V i 



Ex. 143. Diego Ortiz, Tratado de glosas, Rome, 1553, (a) un- 
accented lower note turn, (b) unaccented upper turn : 

Ejioclef In origr] 



Ex. 144. Fray Tomas de Santa Maria, Arte de taner fantasia, 
Valladolid, 1565, unaccented upper turn: 

Ex. 145. Henry Purcell (or his ed.) 3 posthumous Lessons, London, 
1696, unaccented upper turn: 

Ex. 146. C. P. E. Bach, Essay, 1753, II, iv, 23, (a) unaccented 



upper turns on dotted notes (N.B. for rhythm) and at 36, (b) un- 
accented upper turn on undotted note : 



fl.L _ 

rfcn, r rr^g 


Ex. 147. D. G. Turk, Klavierschule, 1789, unaccented upper turns 
(N.B. for rhythm): 

Allegro Moderate Allegro 



(a) The accented upper turn and the accented lower turn both serve 
a harmonic as well as a melodic function, because their initial note, 
which receives the accent, is foreign to the harmony given by their 
main note, and often discordant with it. The case is therefore similar 



to that of the normally prepared baroque trill. A firm attack Is thus 
indicated, though its firmness will be proportionate to the prevailing 

It will be noticed from the examples given above that the accented 
upper turn is the most usual through the greater part of the baroque 

(b) The accented main-note turn and all the unaccented turns are 
primarily melodic, and are more or less unstressed. 

(c) The speed and the rhythm of the turn are both governed entirely 
by circumstances, at the discretion of the performer. 

(d) On dotted notes, the unaccented turn has a strong tendency to 
delay and shorten the note after the dot, unless the tempo is such as 
to make this undesirable. 

(e) A slur, if not written, is implied over the entire ornament, in- 
cluding its main note; except that with unaccented turns on dotted 
notes, the note after the dot may (though it need not) be preceded by 
a brief silence of articulation. 

(f) Accidentals will be found in some of the above examples; but 
their presence in writing cannot be counted upon, even when a sign 
for the turn is marked. Both upper and lower auxiliary notes may be 
diatonic to the key of the passage; either may be chromatic; they will 
seldom both be chromatic at the same time. Nor will they usually both 
be at a semitone interval from their main note at the same time, which 
puts them only a diminished third apart; but this is by no means 
unknown in baroque music, and became less unusual at a later date. 

It will be noticed that an accidental placed above the sign for the 
turn may affect either the upper or the lower auxiliary note. The con- 
vention by which an accidental above the sign affects only the upper 
auxiliary and an accidental below the sign only the lower auxiliary did 
not become established until well into the nineteenth century. 

(g) The interpretation of turns remained unchanged down to and 
including the generation of Beethoven, and very little changed from 
then until the present time. The following remarks are in many respects 
as relevant to the nineteenth century as they are to the baroque period 
at the close of which they were written. 

(280) C. P. E. Bach, Essay, Berlin, 1753, II, iv, 3ff.: 

'The turn is used in slow as well as in quick pieces ... it is not very 
effective over a short note . . . for the most part the turn is performed 
rapidly . . . the turn is better when the next note ascends rather than 
descends ... its sign is little known except on the keyboard. It is often 
marked with the signs of the trill or even the mordent, which two are 
also often confused ... in slow, expressive pieces the turn may give up 
its brilliance and become deliberately languid.' 



C (iv): Broken Chords, C (v): Broken Notes 
and C (vl): Broken Time 


English: Arpeggio 

Italian: Arpeggio 


French: Arpege, harpege 

Arpegement, harpegement 

German: Brechung 



(a) The arpeggio as we now understand it is not regarded as an orna- 
ment. On keyboard instruments and above all on the harp, it is a device 
for spreading the sound, usually beginning with the bass (which should 
normally take the beat) and continuing upwards through one hand, or 
both hands either successively or simultaneously. 

This degree of spreading was so general on the harpsichord, the 
clavichord or the lute in renaissance and baroque music as to have 
needed no special mention. There were also other varieties of spreading 
taken for granted then which could not occur now unless written into 
the notation: from the top downwards (the treble note, especially if it 
is the melody note, taking the beat); from the bottom completely up 
to the top and down again (both this and the former may well occur 
when the melodic line itself continues in the bass); up, down, and up 
again, or up and down twice or more (especially where a long chord 
needs to be sonorously filled out); once up and partly down (to end 
on an inner note from which the melodic line continues); and other 
patterns more elaborate still. Some of the above patterns receive special 
mention in the baroque authorities under the heading of ornaments. 
However elaborate, such patterns are all varieties of the plain arpeggio. 

(b) Where, in addition to spreading the chord in more or less 
elaborate patterns, the performer gives these patterns a melodic value, 



he makes not a plain arpeggio but a figurate arpeggio. He may do so 
exclusively from the notes proper to the chord ; or he may diversify the 
chord by momentarily introducing notes which are foreign to it. The 
notes thus momentarily introduced are acciaccaturas, for which see 
Ch. XVII above. 



In so far as the broken chord is a mere spreading (whether slight or 
considerable) for the sake of enriching the texture and increasing the 
sonority, it arose from the natural technique of such instruments as the 
harpsichord (where an absolutely simultaneous full chord sounds in- 
tolerably harsh), the lute (where the fingers available may not allow it) 
or the lyra da braccio (which is specially designed for partially arpeggi- 
ated chords). 

In so far as the broken chord achieves a truly melodic pattern, it 
arose as part of the general tradition of free ornamentation, selected 
figures from which became to some extent stabilised here as specific 
ornaments in the usual way. 


(a) The toccata in its late renaissance and early baroque sense as a 
'touch-piece 9 for loosening the fingers, establishing the tonality and 
preparing the listener for the more formal music to come was largely 
constructed from improvised arpeggiation. So, too, was the prelude of 
the same periods, before the two forms diverged. The unmeasured 
preludes of the French school of D'Anglebert and others continued to 
be written in a very incomplete form of notation, for the interpretation 
of which some experience with baroque methods of breaking the chord 
is necessary. J. S. Bach's first prelude of the 'Forty-Eight' is a written-out 
example of a series of broken chords such as a majority of baroque 
composers would have notated as plain chords, leaving it to the per- 
former to invent his own figuration. For this reason it was certainly 
meant to be played fast enough and freely enough to give the impression 
of an improvisation, and not in the portentous manner preferred by 
some modern interpreters. On the other hand, it is equally deprived of 
its natural effect if it is played too fast. Between the two extremes there 
lies a rendering which is at the same time relaxed and full of feeling. 

(281) G. Frescobaldi, Toccatas, Rome, 1614, Preface, 3: 

'The openings of the toccatas are to be taken adagio and arpeggiando; 
and the same with suspensions or discords, even in the middle of the 
works, so as not to leave the instrument empty: which breaking is to be 
performed at the discretion of the performer.' 



Ex. 148. G. Frescobaldi, "Toccata ottava 9 , opening bars as inter- 
preted by Arnold Dolmetsch, Interpretation, London [1915] p. 261: 

As written; 

As performed by Dotmetsch: 


Ex. 149. Mace, Mustek's Monument, 1676, arpeggiation of chords, 
transcribed by A. Dolmetsch, Interpretation, London [1915] p. 264: 

As written: 

As interpreted by Mace: 

Ex. 150. Jacques Champion de Chambonnieres, Pieces, Paris, 
1670, broken chords: 

II j jg- 


Ex. 151. Jean-Henri d'Anglebert, Pieces de Clavecin, Paris [1689], 
broken chords: 

eutesur Cheutesur 
unenotft detix noltes 



Ex. 152. J. S. Bach, Forty-Eight, II, Prelude XXI, bar 17, as inter- 
preted by Arnold Dolmetsch, Interpretation, London [1915], pp. 


Interpreted by Dolmetsdt as : 

Ex. 153. C. P. E. Bach, Essay, Berlin, 1753, II, 26, broken 
chords : 

> * 


? ^ 


(b) In virtuoso violin music of the late baroque schools, passages 
occur which are written as chords but intended as arpeggiation. Where 
the composer has not shown the figuration by writing out a bar or two 
in full, the performer has full liberty to introduce what figures he 
prefers, in any variety of rhythm suitable to the context, and with any 
variety of bowing. He is not even obliged to confine himself to the notes 
written in the chords, provided he keeps to the harmony together with 
such extraneous notes as can be momentarily introduced as passing 
acciaccaturas (simultaneous acciaccaturas are confined to keyboard 


D. G. Turk (Klavierschule, 1789) includes arpeggios starting from 
the top as well as from the bottom, with a variety of the usual baroque 
signs for each. But the piano is much less well suited by arpeggiation 
of the baroque kind than the harpsichord. Arpeggio-like written-out 
figuration, meant to be performed with the damper-raising pedal down, 
took its place. 

As late as 1790, however, J. C. F. Rellstab (C. P. E. Bach's Angfangs- 
grunde mit einer Anleitung) published workings of C. P. E. Bach's chord 



passages, with which he took the most striking liberties (examples in 
Doimetsch, Interpretation [1915] pp. 271-2), 

And as late as the first decades of the nineteenth century, J.-B. 
Cartier (1765-1841) was teaching the breaking of violin chords on the 
baroque method, such as the following taken by him from Geminiani's 
Art of Playing on the Violin: 

Ex. 154. Jean-Baptiste Cartier, UArt du violon, Paris [1798], 
broken chords (with a variety of bowings) : 

"' 111 ki 










A note deliberately brought in late for expressive effect (the rest of 
the harmony arriving punctually) was regarded as an ornament by 
some composers of the French school, under the name of suspension 
or demi soupir (compare the medieval hocket). A convenient English 
term for this is truncated note. The Italian word mezzo-sospiro given 
in the following suggests that it may have been originally an Italian 
singers' ornament. 

(282) S. de Brossard, Dictionaire, Paris, 1703, s.v. 'Mezzo-Sospiro' : 

'The figure X * * marks that one is silent for the eighth part of a bar 
[or less or more in proportion].' 

A note of which the latter part is replaced by a silence of articulation 
was similarly regarded as an ornament by some, though in fact this is 
one of the commonest effects of ordinary expression. Its French names 
were aspiration, coup sec, demi soupir, detache, son coupe, suspension, 
tacte, and we can conveniently call it in English a curtailed note. 

Ex. 155. Jean-Philippe Rameau, Pieces de clavecin, Paris, 1731, (a) 
curtailed note, (b) truncated note: 


I have not noticed any post-baroque evidence for either of the above 
regarded as a specific ornament. 



Tempo rubato, besides its normal modern meaning of liberties taken 
with the tempo, had for many baroque authorities the further meaning 
of a displacement of note-values within the bar, or even merely of a 
displacement of the accentuation. There was some tendency to regard 
this as an ornament, and in extreme instances it has certainly an effect 
of free ornamentation; but fundamentally it is a matter of rhythmic 
alteration, for which see Ch. XLI below. 



D* Compound Ornaments 


Two or more ornaments were frequently run together in baroque 
music. When these combinations appear with sufficient regularity to 
become specific ornaments in their own right, it is convenient to call 
them compound ornaments. There is no sharp dividing line between 
these and many others which could be regarded as such with almost as 
much justification; but the following are the commonest combinations. 


(a) We have seen in Ch. XX, 4 above that while all standard baroque 
trills start with their upper (accessory) note, some have this initial note 
so prolonged that it is in fact a long appoggiatura (and was often so 
regarded). There is nothing to be added here except the warning that 
where such an appoggiatura is written out in ordinary notation, it 
should be tied over to include the initial upper note of the trill itself. 
See Ex. 97 in Ch. XX, 4 above, where at (a) and (d) the first note is 
in effect an appoggiatura to the ensuing trill, as opposed to (e), where 
it is not. 

(b) The long appoggiatura from below was so often rounded off 
with a mordent that some authorities made a rule of this, which was 
certainly an exaggeration. 

(283) Jean Rousseau, Traite de la viole, Paris, 1687, p. 87: 

'The Mordent is always inseparable from the Appoggiatura, for the 
Appoggiatura must always be terminated by a Mordent/ 

(284) J. M. Hotteterre, Principes de la flute, Paris, 1707, Ch. VIII: 
'Often one links mordents with appoggiaturas.' 

(284a) Johann Mattheson, Vollkommene Capellmeister, Hamburg, 
1739, Pt II, Ch. 3, 56: 

'In singing, there is hardly an ascending Appoggiatura without a 


(285) C. P. E. Bach, Essay, Berlin, 1753, II, v, 6: 

'When It follows an appoggiatura, the mordent Is taken lightly, In 
accordance with the rule governing the performance of appoggiaturas 
[by which they diminuendo on to their resolutions].' 

Ex. 156. Henry Purcell (or his ed.), Lessons, London, 1696, 'beat* 
(meaning short appoggiatura with mordent): 

Ex. 157. G. T. Muffat, Componinienti, 71736, appoggiaturas with 
and without mordents : 




fl- * or: . W 



(a) The ordinary turned ending which is one of the two standard 
endings of the baroque trill (Ch. XX, 6) was sometimes regarded as a 
trill rounded off with a turn, which is fairly plausible; sometimes as a 
trill rounded off with a mordent, which (because of the rhythm) is not. 
In any case, this is not a genuine compound ornament. 

(b) The ascending trill (Fr. tremblement coule en montant; Ger. 
Triller von unten, Doppelt Cadence) is a genuine compound of ascending 
slide leading to trill; the descending trill (Fr. tremblement coule en 
descendant i Ger. Triller von oben, Doppelt Cadence) is the exact reverse, 
but produced by a turn leading to a trill. 

Ex. 158. J. S. Bach, Clavier-Buchlein, 1720, ascending and descend- 
ing trills: 

Doppelt Cadence Idem 

Doppelt Cadence 
und Mordant 


Ex. 159. F. W. Marpurg, Principes, 1756, Tab. V, ascending and 
descending trills : 

f\ ^-*v- > X\*u .. rHvg0 -.. 

(c) The following are on the borders of the trill and the turn. 

Ex. 160. C. P. E. Bach, Essay, II, iv, 27 and 37, (a) trilled turn 
(prallende Doppelschlag); (b) ascending turn (Doppelschlag von 
unteri) : 

(d) There is a more developed form of the ascending trill to which 
it is convenient to reserve the name double cadence, although this 
served for both in different baroque sources. The ornament became a 
very common cadential formula in the eighteenth century. 

Ex. 161. L'Affilard, Principes, 71694, double cadence ('double 
cadence coupee') (a) as written; (b) as explained; (c) as presumably 

Double cadence coupee 


Ex. 162. Georg Muffat, Florilegium 7, 1695, double cadence: 



Ex. 163. Other possible rhythms for the double cadence: 

(e) Playford (Introd., 1654, 1660 on), Simpson (Division-Viotist, 
1659) and Mace (Mustek's Monument, 1676) agree substantially on an 
English seventeenth century ornament called the double relish, a 
combination of two trills and an appoggiatura from below, on the main 
note or resolution of which a mordent may well be added. The primary 
form is as follows. 

Ex. 164. Chr. Simpson, Division-Violist, London, 1659, double 







Figured Bass 


Until the end of the baroque period it was normal for accompaniment 
to be more or less improvised. From a remote antiquity, the necessary 
understanding between the performers was facilitated by traditional 
modes (in the wide Oriental sense) or other conventions. More recently, 
some partial indication in writing was provided. 

Renaissance accompanists were trained to fill out the harmonies 
from a bass part alone. With polyphonic music, this was not entirely 
reliable. In spite of some useful working rules, it was difficult to see 
quickly enough whether a triad or a chord of the sixth was required, 
whether the third or sixth was major or minor, whether the concord 
was delayed by a suspension and so forth. Many organists went to the 
trouble of copying the parts in score: the first known printed example 
of such a score is Tutti i Madrigali di Cipriano di Rore a 4 voci, spartiti 
('divided', i.e. barred) et accommodati per sonar d'ogni sorte d'istru- 
mento perfetto . . ., Venice, 1577. At this time, the ideal standard 
accompaniment was regarded as one which reduplicated the written 
vocal parts; and this the score made easy. A close approximation could 
also be effected where the bass line was copied or printed with the top 
line of the polyphony shown above it. 

Where the polyphony is distributed among more choirs than one, 
each has its own bass line, but only one of these can form the true bass 
of the harmony at any given moment. In some organ parts of the late 
sixteenth century, the several bass lines are printed, it being left to the 
organist to pick out the true bass; in others, not the bass lines but the 
true bass extracted from them is printed. This came to be called a 
Thorough or Through Bass, Basso Continue*, Generalbass or their 
equivalents, meaning a bass continued throughout the music by taking 
whatever happened to be the lowest part for the time being. And 



presently (perhaps first in Viadana's Cento concert! ecclesiastic^ 
Venice, 1602) portions or all of this continued bass were not merely 
extracted from the parts to be accompanied, but added to them. 

A few particularly ambiguous bass notes now had f or b above them 
to show the major or minor third. These are the 4 signa' or signs, which 
soon included 'numeri' or figures (the first being a 6); from these 
developed the baroque system of Figured Bass, Basso numerato, Basse 
chiffree, bezifferter Bass or their equivalents. 


A figured bass is thus a thorough-bass or continue to which acci- 
dentals and figures have been added showing the main intervals 
required. But the distribution of the intervals, the conduct of the parts 
and the melodic figuration are not shown. The performer Is told, with 
very varying degrees of thoroughness and accuracy, what harmonies 
to produce; how he produces them Is his own affair. 

There are some obvious practical advantages in this ingenious form 
of musical shorthand. There is first the great saving in composing and 
copying time and printing costs; there is next the flexibility of an 
arrangement which enables each performer to adapt his accompaniment 
to the requirements of his own instrument. A lutenist can employ his 
typically thin and widely spaced distribution of notes: sketchy to the 
eye; unexpectedly sonorous in performance, and suggestive of more 
complex part-writing than is actually present. A harpsichordist can fill 
out his instrument with resounding arpeggios and acciaccaturas; an 
organist can sustain counterpoint pure enough and sparse enough not 
to overload the texture. Any fully written-out part which suited one 
could only hamper the others; but a figured bass part hampers nobody. 
It leaves the crucial decisions of sonority and spacing to the circum- 
stances of the moment: the instrument; the resonance of the hall; the 
balance of the ensemble as a whole. 

There was also a more fundamental consideration. It was not from 
mere casualness that baroque composers preferred to trust their per- 
formers with figured bass, as they also trusted them with free orna- 
mentation. It was from the high value they set on spontaneity. They 
believed, and modern experience confirms, that it is better to be accom- 
panied with buoyancy than with polished workmanship. An accom- 
panist who can give rhythmic impetus to his part, adapt it to the 
momentary requirements of balance and sonority, thicken it here and 
thin it there, and keep every bar alive, can stimulate his colleagues and 
help to carry the entire ensemble along. This is not merely to fill in the 
harmony; nor is it merely to make the harmony into an interesting 
part; it is to share in the creative urgency of the actual performance. 



As with free ornamentation, It hardly matters how much is actually 
improvised, and how much is memorised or written out by the per- 
former or the editor. It only matters that it should have the fresh and 
flexible feeling of an improvisation. Some movements lend themselves 
readily to an improvised realisation; others present opportunities of 
which the fullest advantage can only be taken by working out a part 
on paper. Such a worked-out part, however, can itself often be improved 
in performance, and may not reach its best until it has been played 
many times. 

Almost all the more old-fashioned modern editions have realisations 
of the figured bass which are so bad that they really spoil the music 
heavy, pianistic, and with a strangely nineteenth-century flavour to the 
harmonlsation even when they are not palpably contravening the 
figuring. Recent editions are generally quite usable, but they still suffer 
from two very common faults: the part is nearly always too thick and 
needs to have a great many notes thinned out before it will take on a 
natural, easy flow; and at the same time the real musical interest is too 
slight. Such a part is at once too much and too little. There is too much 
sound for a clear texture, but too little musical significance to match 
up to the solo melody. The present tendency is a great relief from the 
previous romantic over-elaboration, but it has on the whole swung too 
far in the direction of dull correctness and austerity. The balance, 
indeed, is not an easy one to strike; and it must also be remembered 
that there is no one right solution. The essence of the system is that it 
leaves the performer free to find his own solution within the given 
indications and the general boundaries of the style. 

All this points to the desirability, for any accompanist seriously 
interested in baroque music, of learning figured bass: if possible to the 
extent of realising his own parts; and at least to the extent of being 
able to modify an editor's parts to his own taste, which makes a very 
good start. No realisation, after all, is sacrosanct: whoever made it, 
and however well, it is only one working version out of the unlimited 
possibilities. That is the best of figured bass. 

The following treatment is confined to the essentials, but F. T. 
Arnold's lengthy Art of Accompaniment from a Thorough-Bass (London, 
1931) is one of the few entirely satisfactory modern studies in early 
interpretation, and anyone going far into the subject will certainly turn 
to it for its mass of detailed and accurate information. Some knowledge 
of harmony and counterpoint is a necessary prerequisite; I know of no 
better introduction to these than EL K. Andrews' articles on them in 
Grove's Dictionary (5th ed., London, 1954). 



The Figuring 


It is in the nature of baroque music to rest on a bass of which the 
outline is firmly melodic. This bass is the foundation on which the 
continuo player will construct his accompaniment, and is always given 
in his part complete. 

A minority of baroque musicians continued to regard this bass line 
as a sufficient indication in itself, without figures, either because they 
expected the accompanist to make a short score or tablature and 
play from that, or because they thought figures were an unnecessary 
complication for a competent accompanist. 

(286) Lodovico Viadana, Concern Ecdesiastici, Venice, 1602, pre- 
face, rule 6 : 

1 hope that the Organists will be able to make the said tablature at 
their own convenience, which, to tell die truth, is much better/ 

(287) Giovanni Piccioni, Concern Ecclesiastic^ Venice, 1610, 
preface : 

1 have prefered not to set any sort of signs, such as sharps, flats, and 
figures, over the notes, as many do, since to unskilful Organists, they are 
confusing rather than otherwise, while to the knowledgable and expert, 
such signs are unnecessary, for they play them correctly by ear and by 

A similar attitude persisted in some quarters even into the eighteenth 

(288) Andreas Werckmeister, Harmonologia Musica, Frankfurt and 
Leipzig, 1702, sect. 120, p. 65: 

'Those, again, who say that the signs above the notes in a continuo 
bass are perfectly useless and unnecessary, show no little ignorance and 
stupidity: for it is plainly impossible even for an expert who knows the 
natural movement of harmony and composition to perform everything 
correctly in accordance with another man's ideas, since the progressions 
and the resolutions may take many different courses.' 

(289) C. P. E. Bach, Essay, II, Berlin, 1762, XXXV, 2: 

'Some people have been at great pains to elaborate a system for realising 
unfigured basses, and I must confess that I have made experiments of the 



kind. But the more thought I have given to it, the more richness I have 
found in the turns of harmony ... it is impossible to devise fixed rules. 
[3:] an accompaniment made from an unfigured part can only be inade- 
quate. . . . He who is too lazy or too ignorant to figure his own basses 
as a good performance requires should get it done for him by an expert 

It must be admitted that the necessity for figures is somewhat less 
pressing when the progressions are very regular and largely comprised 
of sequences. It is interesting that the Turin MSS. of Vivaldi's music 
often give no figuring where the printed editions have figuring supplied, 
not very well and presumably by the publishers. Moreover, when at 
least one melodic part is printed above the bass, the need for figuring 
is much reduced. But though rare exceptions occur in which more than 
one melodic part is shown in this way, it is ordinarily only in the case 
of solos that this is done. 

The best course for a modern accompanist who wants to realise an 
unfigured bass at sight (especially where no melodic part is shown 
above the bass) is to look at the other parts and put in sufficient figuring 
for his own information. 


(a) Ordinary numerals show the intervals of the required harmony, 
as reckoned from the bass. These intervals are shown simple, without 
regard to octave transposition. Thus 3 implies the interval of a third, 
a tenth or any further octave transposition required. The 'position' of 
each chord is, of course, defined by which note of it forms the bass; 
but above that, its notes may be spaced in any order desired. 

This rule applies to all figured basses after Peri, Caccini, and 
Cavalieri, with only one partial exception of any importance : 2 and 9 
are figured separately, because of the different progressions to which 
the intervals of a second and of a ninth incline. The distinction, how- 
ever, was not always kept free from confusion. Moreover, it must be 
realised that even with these two intervals, octave transposition is still 
taken for granted. An interval correctly figured 2 may still be taken an 
octave and a tone, or even two octaves and a tone, above its bass note. 

(b) The numerals shown are normally those regarded as sufficient to 
define the chord in question. Further intervals needed to complete the 
chord were assumed, and were only figured for some special reason, 
usually a desire to indicate the actual progressions between the chords 
(i.e. the Voice-leading'). 

This is a very important point in practice. Any unnecessary figures 
make the part much harder to read. Complete figuring would merely 



be a less efficient way of writing out the realisation; the whole point of 
figured bass is that, being a form of shorthand, it makes a fully written- 
out part superfluous. 

The art of reading from a figured bass largely depends on having 
memorised all the combinations of figures normally encountered, 
together with the unfigured intervals which these combinations imply 
to complete the chord in question. This memory becomes so quick that 
it almost appears to be localised in the fingers themselves, which can 
be trained to take up the spacing required by the pattern of each chord 
without the need for conscious thought for which there is no time. 

Circumstances may call for any feasible amount of doubling. Not all 
intervals in a chord can be doubled with equally satisfactory effect, and 
there were a number of rules and conventions to give the accompanist 
guidance in this matter. These varied in detail in different periods, 
schools and authorities, but were basically in accordance with the 
ordinary grammar of composition as it is taught in our better colleges 
of music and university music departments today. For a thorough 
treatment both of the basic principles and of the variations of detail, 
the reader is referred to Arnold (op. tit.). The same applies to the 
principles and details of 'voice-leading* by which the progression from 
chord to chord is actually effected: the basis is our normal academic 
basis, and the variations are best studied in Arnold. 

Not only doubling of intervals but thinning out may be called for 
by circumstances. Three-part or even two-part harmony sometimes 
meets the momentary requirements of sonority and texture better than 
four-part; very light movements may benefit from such thinness of 
harmony throughout. The choice of the best interval or intervals to 
omit is less subject to rule than the choice of intervals for doubling, 
but some brief guidance is given below, and more will be found in 
Arnold (pp. cit.). 

The intervals which, subject to any deliberate doubling or thinning, 
the figures imply in addition to those which they themselves indicate, 
are the subject of very definite conventions, to which attention will be 
drawn here under each chord in turn. 

(c) Accidental signs have the same effect on the intervals shown by 
the figures against which they appear as they have on ordinary notes 
in the stave. 

But it is important to remember the differences between baroque and 
modem conventions with regard to accidentals, for which see Ch. IV, 
4, Ch. V generally and Ch. VI, 4, 6 and 9. 

In particular, the need to sharpen leading notes and to make major 
(Picardy) thirds on important closes, whether so written or not, must 
be taken into account; individual cases are not always easy to judge, 



especially at sight, but both these are conventions which an accompanist 
cannot altogether ignore in baroque music of any date. 

As usual in baroque music, accidental signs sometimes show a 
specific pitch, as they do in modern times, but sometimes state a 
chromatic modification, or merely a chromatic situation, irrespective 
of specific pitch. Thus, for example, a b 5 is often used to indicate a 
diminished fifth even where diatonic to the key. This does not mean 
that the 5 is to be chromatically flattened; it merely warns the reader 
that he must play a fifth which is flatter by a semitone than the one 
he might otherwise assume: i.e. a perfect fifth (see 4 below). On a B tj 
bass he would normally play F# if the D were shown sharp, and might 
do so even if it were not; hence the figuring for an F natural, even in 
the key of C where it is diatonic, might either be 5 (as we should now- 
adays expect) or b 5 (as a precautionary measure to make sure of it). 
No baroque performer could mistake that foi an Fflat, which a modern 
performer might conceivably do for a moment. 

(d) The figures themselves may appear in special forms indicative of 
chromatic modification and cancellation. 


A. Figures in a form equivalent to an added sharp 
t 24t?&7$ * 

2 f 5f 9- 

B. Figures in a form equivalent to an added natural 

C. Figures in a form equivalent to an added flat 
4fc& & T*l*l 


(e) A sign - placed over the figures is occasionally met. It was the 
invention of G. P. Telemann, and excludes some interval from the 
chord which the figures would normally imply. 

(f ) A more or less horizontal dash or line following a figure prolongs 
the note given by this figure until the dash or line ends; unfortunately 
this only became established in the latter part of the baroque period. 

(g) Most composers were to some extent illogical and inconsistent in 
their notation of figures and other signs, and the performer must not 
be deceived by apparent absurdities but use his intelligence to infer the 
actual meaning. Some (e.g. Abel, Leclair) were definitely idiosyncratic 
on occasion: Arnold gives examples (op. cit.). 




(a) When a bass is figured, the figures are normally placed above the 
stave carrying the bass line. But they may also be placed below, with 
identical effect. 

(b) When an accidental is added to a figure, it may be placed either 
before or after the figure, with identical effect. There was little con- 
sistency in this matter, except that the accidental might tend to be placed 
on the side where there was most room for it when space was short. 

(c) When more harmonies than one are to be played in succession 
on a single bass note, the exact horizontal spacing of figures is supposed 
in theory to give some indication of the timing of the subsequent 
harmonies; but in practice it hardly ever does. It is very common to 
find all the figures bunched together towards the left side of the bar at 
the beginning of the bass note, with no means of seeing what is meant 
to be the duration of the harmonies thus figured. Even if the figures 
are spaced out, the spacing is only approximate no copyist or com- 
positor could be expected to do more than that, if as much. 

The earliest figured basses of the monodic school (Peri, Caccini, 
Cavalieri) met this difficulty by dividing a long bass note into as many 
shorter notes as there are changes of harmony, each such short note 
being of the exact duration required, and bearing its own figures; and 
a tie being placed over the whole to show that these short notes are not 
played themselves as separate notes, but are one long note notated in 
this way to give the necessary information. This method was apparently 
regarded as too clumsy to be continued; but it had the advantage of 
complete unambiguity. Incidentally, to a string player this notation is 
misleading in another way: it looks like aportato, i.e. a slight reiteration 
of emphasis within one bow. He must be warned that there was no 
such intention here. 

(290) Giulio Caccini, Nuove Musiche, Florence, 1602: 

'I have used the ties in the bass part because, after the consonance [or 
other interval figured] only the interval figured is to be struck again [and 
not the bass note].' 

Ex. 165. Caccini, method of showing time values of figures: 

1 10 11 11 #10 14 

m ~ 

Here the bass note A is held for a semibreve; but the chords are held 
for the time values which the tied subdivisions of the A depict: i.e. 
three crotchets and two quavers. 



In the main baroque period, conventions grew up which, though less 
reliable, gave some assistance. 
(291) C. P. E. Bach, Essay, II, Berlin, 1762, 1, 43-7: 

'Figures placed over the dots which prolong notes are played at the 
time when the dot comes into effect, though they relate to the previous 
note. . . . Figures placed over short rests are played at the rest, but relate 
to the following note. . . . Figures over long rests are also played at the 
rest, but relate to the previous note. ... A trained musician can readily 
decide which is in question by looking at the context. 

'Figures placed after a [duple] note are played on its second half ... if 
two figures are placed over a duple bass note, each figure is given half of 
its duration. . . . 

c lf three figures in succession are placed over such a note, the first 
figure, immediately above the note, is given half its length and the 
others take half each of the remaining half. . . . If two figures in succes- 
sion are placed over a dotted note . . . the first figure is given two- 
thirds . . . and the second the remaining third. ... If there are three 
figures in succession, each is given a third of the duration.' 

It was also possible to mark a short dash after a figure to prolong 
it at the expense of the others. But in practice, it is most often necessary 
for the performer to discover the requisite distribution with little or no 
visual aid; and this is, in fact, one of the chief details to look out for 
if it is possible to take a brief, but only a brief glance at the music 
before realising the part at sight. 

(d) The normal vertical arrangement of the figures is to place those 
of higher denomination above those of lower denomination. When they 
are disposed in a different order, this may be to show a particular 
spacing of the harmony, or it may be without notational significance. 

(e) In very early figured basses, an accidental set a little to the left 
of a bass note, and a third above, was intended to show a modification 
of the third to major or minor as the case may be. An accidental 
similarly set a sixth above or (more likely) a third below, was intended 
to show a like modification of the sixth. These accidentals are very 
frequently misplaced by copyist or printer. When, in this music, an 
improbable accidental appears at the same pitch as the bass note, and 
with apparent reference to that note, it is worth asking whether the 
intention was to modify not the bass note but either the third or the 
sixth of the chord, according to circumstances. 

Later in the seventeenth century, cases occur of accidentals placed 
above or below the figures to which they apply, instead of before or 

(f) A rare convention, but troublesome enough to be worth noting, 
can only be called (as it is by Arnold, op. cit. y p. 880) retrospective 



figuring. It occurs on a bass moving by a step, the second note of which 
bears figures reckoned not from that note., but from the previous note. 

Ex. 166. H. Purcell, Sonnattfs of III Parts, 1683, Sonata VII, 1st 
movt, bar 9, retrospective figuring as realised by Arnold, Thorough- 
Bass, London, 1931, p. 880: 



It will be seen that the figures are as if the bass had been a semibreve 
F sharp, and that this would have made a conventional cadential 
formula of which the actual notes are merely a slight variant. This is 
the best explanation I can see; but it is more understandable than 
logical, and is an excellent example of the perennial need to pick out 
the correct solution from the literal absurdity. The logical and proper 
fingering of the E bass note is, of course, # 4-2. 


(a) If a bass note is essential (and not, e.g., merely a passing note, 
for which see 14 below), it will normally be given fresh harmony, 
whether figured or not. 

In the absence of any figure, this harmony is normally assumed to 
be a diatonic triad. 

(b) But in all except the most amply figured basses, this assumption 
is subject to exceptions. Any bass note chromatically sharpened to 
become a leading note is assumed, in the absence of figuring, to bear 
not a triad but a chord of the sixth. This is often extended to leading 
notes in the bass even when diatonic. And in general, in any context in 
which to a trained accompanist the choice would be fairly obvious, we 
are liable to find that an unfigured bass note requires not a triad but a 
chord of the sixth. This is one of the many matters in which a good 
accompanist needs an extra sense to guide him; and it must be admitted 
that in this respect it is not particularly difficult to acquire. Genuine 
ambiguity, even when sight-reading, between triads and chords of the 
sixth in the absence of figuring is much less frequent than might be 
expected. Generally the progression speaks for itself, given familiarity 
with the style. 

(c) In very ambivalent cases, the correct figuring to prevent mistakes 
is 3. This states that the third is present. 

The presence of the third, whether assumed or indicated, implies the 



fifth to complete the triad, in the absence of any further assumption or 
indication to the contrary. 

The figure 3 also occurs to indicate the resolution of a 4 preceding 
it on the previous portion of the same bass note (or its octave), or in 
other situations where it is desired to draw particular attention to the 
third. Other intervals may also be figured for similar reasons (i.e. the 
5 and the 8) either above or in combination. 

But where a third requires chromatic modification, the necessary $ 
or b or t] is normally set in the position which the 3 would otherwise 
occupy, the 3 itself being omitted. Much less commonly, both accidental 
and figure are shown, as $ 3 or b 3 or b| 3. 

The sign t|, however, remained comparatively infrequent in the 
figuring of basses (more so than in the rest of the text) and the reader 
must be prepared to find # as the cancellation of b and b as the cancella- 
tion of $ (see Ch. IV, 6 above). 

When the 5 but not the 3 is to be accidentally sharpened, the figuring 
* (or # 5, etc.) is necessary (for variants, see Tables II and III in 
Ch. IV, 4 above). 

When both 3 and 5 are to be accidentally sharpened, the figuring $ 
or iq is necessary, there being no need to indicate the sharpened fifth, 
because in the absence of indication the 5 with a major 3 is normally 

assumed to be perfect, not diminished. Nevertheless the sharpened 5 is 

* g* 
sometimes shown, the figuring then being ^ or , as the case may be. 

?r H 

The same assumption (that 5 with major 3 is perfect) operates when 
the bass is accidentally flattened, so that no figuring is then needed. 

(d) It is normally desirable (except in sequences) to avoid doubling 
the major third in any context which makes it function as a leading 
note (especially if it has been chromatically altered to serve that 

It must be realised that for a leading-note acting as such (in whatever 
inversion or part, of whatever chord) there is only one alternative 
to its natural progression by a semitone upwards. This alternative is a 
drop of a major third to the fifth of the ensuing tonic, which is allow- 
able provided the leading-note is not in the top part (cases abound in 
J. S. Bach's four-part chorales). Even in a free style, these rules should 
command respect; they are deeply rooted in acoustic reality. 

Otherwise, any intervals in a diatonic triad, major or minor, can be 
doubled in any suitable way. 

(e) When thinning the chord to three parts, the most natural note 
to drop is the 8, and thereafter the 5. In two parts, the 3 itself must 
sometimes be dropped in the interests of good part-writing, and also 
of harmonic variety. 




(a) No figuring is needed for diatonic diminished triads (comprised 

of minor 3 plus diminished 5), but 5 or 5 or 5 b may be found. The 
semicircle over the figure is to exclude the 6 which might otherwise be 
assumed, although not figured, since a diminished triad and a 6-5 chord 
are In very many contexts interchangeable, and the fuller chord would 
generally be preferred by most accompanists, other things being equal. 

5 b or 5 kj is usually a sufficient indication for a chromatically reached 
diminished triad, unless the third is also modified, when it should be 

The semicircle, being rare except in text-books, may not be present: 
j? 5 or t| 5 is the standard indication. 

It will be noticed that the b or tj to the 5, if found with diatonic 
diminished triads, merely indicates a diminished fifth, but if found 
with chromatic diminished triads, indicates the chromatic flattening of 
the fifth. 

(b) The best interval to double depends on circumstances. When the 
chord stands on the seventh degree of a major scale or the sharpened 
seventh degree of a minor scale, the best (and in pure style, the only 
correct) interval to double is the 3. In its only other standard situation, 
which is on the second degree of a minor scale, the 3 is better than the 
8 to double, and the 8 is better than the (diminished) 5; but the part- 
writing will be the main consideration. The truth is that this is essentially 
a three-part chord, which is never at its best in four-part writing. 

(c) Thinning to two parts can only be by dropping the 3, since the 
diminished 5 is essential to the chord. 


(a) Augmented triads (comprised of major 3 plus augmented 5) are 
figured $* (or # 5, etc.) or t) 5 as the case may be. 

(b) It is most inadvisable (and in pure style, impossible) to double 
the augmented 5, which needs to progress by a semitone upwards and 
cannot properly take any other course. This chord, like the diminished 
triad, is at its most natural in three parts, as C. P. E. Bach points out 
(Essay, Part II, 1762, Ch. 5, 6). 

(c) Thinning to two parts can only be by dropping the 3; but the 
chord is most unsuited to two-part writing. 


(a) Six-three chords are the first inversions of triads. The figuring is 
6 if diatonic; & (or # 6, etc.), t) 6, or b 6, as the case may be, if 



The 3 is implied, but not normally figured unless requiring accidental 
modification, when the # or t; or b is shown, but not usually the 3 itself. 

Fuller figuring is sometimes given to show the progression from the 
previous harmony. 

(b) When the bass (representing the 3 of the triad) functions as a 
leading note (especially if accidentally modified for the purpose) it is 
frequently left unfigured, the 6 being assumed. 

(c) When serving as leading note (especially if accidentally sharp- 
ened), the bass is not normally a desirable interval to double (except in 

If the triad inverted is either diminished or augmented, the note now 
representing the leading note is unsuitable for doubling (and so is that 
now representing the augmented 5). 

(d) The 3 can be dropped in two-parts, but the chord, as such, 
ceases to exist. 

(e) Sequences of six-three chords are natural and admirable in three- 
part writing. In four-part writing, they need some care to make the 
progressions reasonably correct (see Arnold, op. cit., pp. 525 ff.). The 
fourth part can best leap, alternately by contrary and direct motion. 

The 3 should not be above the 6 in consecutive six-three chords, or 
parallel fifths will result. These may be tolerable, however, if one of 
them is diminished. 


(a) Six-four chords are the second inversions of triads. The figuring 
is 4 if diatonic, with the appropriate accidentals for any chromatic 
modifications required by modulation, or by any diminished or aug- 
mented interval in the triad inverted. 

(b) The 4 is strongly inclined (and in pure style compelled) to pro- 
gress by step downwards (in pure style, it also requires preparation if 
on a strong beat). 

But if the 4 remains stationary from the previous chord and into the 
next chord, or if the previous chord was another inversion of the same 
chord, or if the chord is taken as a passing chord on the same bass, 
the above does not apply. 

(c) Doubling is reasonably free except for the inverted leading note 
when the triad inverted is an augmented one. 

(d) In two parts the 4 must be dropped; but the chord, as such, 
ceases to exist. 


(a) Chords of the augmented sixth consist of an augmented sixth 
with a major third ('Italian sixth'); with a major third and an 



augmented fourth (Trench sixth'); or a major third and a perfect 
fifth ('German sixth'). 

(b) The 'Italian sixth' is figured & (# 6, etc.). The proper figuring for 

the Trench sixth' is 4 and for the 'German sixth' - 

(c) The augmented 6 requires no preparation, but should resolve by 
the step of a semitone upwards. The bass moves a semitone downwards, 
and the 3 should do likewise. The 4 of the Trench sixth' remains 
stationary into the next chord. The 5 of the 'German sixth' is better if 
prepared; it should resolve a semitone downwards with the bass, the 
resultant parallel fifths being acceptable in a free style (though some 
authorities both contemporary and modem would deny this) provided 
they are neither between outer parts, nor between tenor and bass in 
other words, provided only one of them is on the outside, and they are 
also separated from one another by an intervening part. 

In most cases, however, these parallel fifths will have been averted 
by the composer. He will have figured a 4 suspension on the note to 
which the bass moves, followed by a | resolution on the same bass note. 

(d) No notes can be doubled without some risk of introducing either 
consecutives or faulty progressions. 

(e) No notes can be dropped without changing the identity of these 
chords, except that in theory the 'Italian sixth' might exist in two parts 
without its 3 if ever occasion for it arose in practice. 


(a) Chords of the seventh are normally indicated by 7 alone, except 
when the third needs accidental modification, and this is shown as 
usual. Fuller figuring occurs when the progressions are to be indicated, 
or when the presence of the 5 or the 8 is to be indicated. The presence 
of a 3 not accidentally modified may indicate that the 5 is to be omitted. 

The 7 is a discord, whose normal resolution is by step downwards. 
In the pure style, it needs preparation; and even in a freer style, it 
should be prepared in most ordinary contexts unless there is a good 
musical reason to the contrary. But except in the pure style dominant 
sevenths and diminished sevenths may be taken unprepared without 
hesitation, and so may chains of sevenths resolving from one on to the 
next. The resolution of the seventh may also be suppressed or omitted 
in the free style, if there is good musical reason, particularly when the 
7 remains stationary into the next chord, or when the bass moves to 
the first inversion of the same chord, or when the bass moves to what 
would have been the note of resolution (transferred resolution), or the 
whole chord moves to one of its own inversions. 

Though there are plenty of exceptions, it is generally undesirable to 



sound the note on which any discord is going to resolve in the same 
chord in which the discord itself is sounded (but see Ch. V, 5 above). 
The effect is not only apt to be muddy, but tends to anticlimax. 

(b) The 3 as leading note should not normally be doubled, except in 
sequences. Nor, in general, should the 7. 

(c) The most natural interval to drop when thinning chords of the 
seventh to three parts is the 5, whose presence is not by any means 
always desirable even in four parts (it may be better to include the 8 
instead, which keeps the chord considerably less weighty). 

In two parts, the 5 and the 3 must be dropped; but though quite 
possible in this form, the chord of the seventh becomes too empty for 
very much use in two-part writing. 


(a) Six-five chords are first inversions of chords of the seventh. The 
figuring is 5, together with any necessary accidental modification to the 
3, shown as usual by accidentals normally without the figure 3 itself. 

The dissonance here is the 5, which is the 7 inverted, and is subject 
to similar stipulations with regard to preparation and resolution. 

(b) The bass as leading note should not normally be doubled, except 
in sequences. 

(c) In thinning to three parts, the 3 should be dropped. The chord 
cannot exist in two parts. 


(a) Four-three chords are second inversions of chords of the seventh. 
The figuring is 3, the 6 being added only when accidentally modified, 
or when it is inserted to show the progression. 

On the supertonic of a prevailing major or minor key, and on the 
submediant of a major and sometimes of a minor key, the single figure 
6 was conventionally held to imply 4 as well as 3 at the discretion of 
the performer i.e. a four-three chord in place of a six-three. 

The dissonance here is the 3, which is the 7 of the uninverted chord, 
and is subject to similar stipulations with regard to preparation and 

(b) The 6 as leading note should not normally be doubled, except in 

(c) The note to be dropped in three parts is either the 6, or when the 
6 cannot be spared because it is functioning as leading-note, the 4, 
though this unfortunately robs the discord of actuality. The chord 
cannot exist in two parts. 




(a) Four-two chords are third Inversions of chords of the seventh. 
The figuring is normally |, which is frequently abbreviated to 2, and 
sometimes expanded to include the 6. 

On the subdominant of a prevailing major or minor key, the 4 is 
augmented, and a frequent figuring is 4t (or $ 4, etc.) or t] 4 as the 
case may be. 

When the second is augmented, the most frequent figuring is f 2 or 
b| 2, as the case may be, or 2$. serving as the equivalent of either. The 
fourth is then augmented too, but is nevertheless not usually shown in 
the figuring, even when accidental, the presence of the augmented 2 
being normally regarded as implying the augmented 4. 

The dissonance here is the bass note, so that its preparation, if any, 
and its resolution, which is normally indispensable, are the responsi- 
bility of the composer and not of the performer. 

(b) The augmented fourth as leading note should not normally be 
doubled, except in sequences. 

(c) In thinning to three parts, the 6 should be dropped; in two parts, 
the 6 and the 4. 


(a) Passing notes in the normal sense are notes joining harmony 
notes by step; but from the viewpoint of figured bass any note which 
is passed over, in the sense of not bearing harmony of its own, can be 
called a 'passing note' even if it leaps (and in whatever direction). 

To distinguish 'passing notes' in the bass, which merely require the 
previous harmony to be held on, from harmony notes in the bass, 
which require their own appropriate chord to be struck whether shown 
by figures or not, is one of the chief practical problems in realising a 
figured bass at sight. When no upper part is given, the problem cannot 
invariably be solved with certainty; even with a good knowledge of the 
style (which is the main help here) mistakes are virtually unavoidable 
unless the composer has gone to unusual lengths to impart the necessary 
information. But in working on paper with the other parts available to 
refer to, the problem is no longer troublesome, since it is always 
possible and seldom difficult to find a realisation which fits the existing 

(b) When the passing note in the bass is unaccented, the harmony 
of the previous note is continued. This may be shown by the very 
simple and satisfactory device of a dash over the passing note. The 
same applies when there are more passing notes than one, in which 
case the dash is prolonged across all of them. But unfortunately the 



dash did not come into common use until the end of the seventeenth 
century, and even then was not employed with the regularity it deserved. 
Occasionally the passing note is itself figured to show the harmony 
carried over from the previous note (rather than struck again). The 
cases, however, in which no direct indication is given are by far the 
most frequent. 

A special case of a 'passing' note in the bass which requires knowing 
is shown by the following example. The crotchet F, being unfigured but 
approached by leap, looks as if it should have its own harmony, i.e. a 
triad, but is in fact treated as though it were a passing note from a 
(non-existent) E. 

Ex. 167. J. D. Heinichen, General-Bass, Dresden, 1728, Part I, 
Ch. IV, sect. 33, p. 323, irregular passing note in the bass (the F in 
bar 2): 

(c) When the passing note in the bass is accented, it carries the 
harmony of the next note, and is in effect a measured appoggiatura (or 
unprepared suspension) in the bass. Perhaps the commonest way of 
showing this is to place the figuring over the second, unaccented note, 
leaving it to the performer to understand that he has to strike it over 
the first, accented note. An improvement on this by no means self- 
evident method was to place a dash over the first note, connecting it 
with the figures shown over the second note. Better still, but by no 
means as common as it should have been, is the correct figuring for 
the passing note placed over that note, and held on through the next 
(and actual harmony) note. Here, too, it is common to find no direct 
indication at all. One indirect indication of the harmony intended may 
be the necessity for it in order to prepare an ensuing discord. Another 
moderately dependable clue is the presence of a leap following the 
short notes in question. 

(d) With both unaccented and accented passing notes in the bass, 
the most decisive factor is their speed. In common \ time, the crotchet 
is the standard unit of harmony; shorter notes in the bass are apt to 
be passing notes unless otherwise figured. In allabreve \ time, the 
minim is the standard unit of harmony, or under some circumstances 
in the pure style, even the semibreve; shorter notes in the bass again 
tending to be passing notes. But in what was sometimes called the semi- 
allabreve time (in effect a rapid \, not a steady |) the crotchet is the 



standard unit; and there are other complications: In \ the minim, in 
\ the crotchet, in | the quaver is the unit, but either one or two of these 
may go to a single harmony. 

No great reliance, however, should be placed on these and other 
baroque rules of the kind. Still less reliance should be placed on the 
various non-numerical signs, C, (J, etc., since these were in a state of 
utter confusion throughout the baroque period, as the best authorities 
of that period unhesitatingly admitted (Ch. XXXVIII below). The per- 
former should allow his own musicianship to tell him, irrespective of 
the signs, what is the main harmonic pulse of the piece in question. 
Notes in the bass shorter than this pulse will tend to be passing notes 
unless otherwise figured. There will be very numerous exceptions, many 
of them undetectable except by the instinct bom of long experience or 
by reference to the other written parts. 

When an accompanist who is realising a figured bass at sight feels 
completely doubtful, he has a choice of evils. He can evade the diffi- 
culty by a momentary gap in his harmony. He can treat the ambiguous 
bass note as a passing note in the knowledge that if this is a mistake, 
it will be a less conspicuous one than the opposite mistake of striking 
a fresh harmony on what was meant to be a passing note. He can, very 
often, thin his part down to tenths moving in parallel with the bass, 
which will in many cases sound right on either basis. 

On the other hand, if he is experienced, the occasions on which he is 
seriously in doubt will be far less frequent than might be expected in 
theory. The harmonic movement is much more foreseeable in some 
styles than it is in others, but even in the least foreseeable there are 
grades of probability to which an experienced accompanist becomes 
remarkably sensitive, not only over this particular problem, but over 
many others. It is only an occasional situation which should be capable 
of defeating him completely. 

It must be admitted that such situations do occur. 



Going Beyond the Figures 


Figurings, like signs for ornaments, are not commands but hints. They 
give information about the harmonies going on in the written parts to 
which the accompanist has to fit. They may also impart the composer's 
wishes; but this is not usually their main function. 

So long as polyphonic parts were regarded as best accompanied by 
doubling them, the two functions coincided. 

(292) Bernardo Strozzi, 'Concerti Ecclesiastic! 1 , III, preface (quoted 
with approval by Praetorius, Syntagma, III, 1619, Ch. VI): 

* A great deal of discordance was heard [with unfigured basses] because 
everyone applies the rules of music according to his own methods, whim, 
disposition and fancy . . . [Figures were invented as a means by which] 
a man might play it correctly, in such a way that no mistakes were heard, 
and might handle it as far as possible in the manner intended by the 

'With the help and use of the said figures, they handled and performed 
the Motets of Palestrina ... in such a way that it appeared to the listeners 
exactly as if the whole had been set out in full Tablature.' 

The Italian monodists, who took up the use of figured bass on its 
first appearance and were certainly among the pioneers in developing 
it, by their use of compound figuring controlled the progression of their 
harmonies to an extent impossible so soon as compound figuring was 
dropped in favour of simple figuring regardless of octave transposition. 
This change itself is evidence of a change of policy. Before many years 
had passed, we have clear evidence for what became the accepted ideal 
of standard figured bass accompaniment: a more or less chordal 
texture against which the written parts can stand out, as opposed to 
being doubled. Ironically enough, the basis of this ideal was the part- 
writing of the great polyphonists, which is the starting point for good 
harmonic progression in the chordal style. 

(293) Johann Staden, Kurzer Bericht, Nuremberg, 1626: 

'[It is] good authors such as Orlando [di Lasso], Luca Marenzio, 
Claudio Merulo, Pierluigi Palestrina, Andrea and Giovanni Gabrieli, 
Horatio Vecchi, Giovanni Paulo Cima, Johann Leo Hassler, Gregor 
Aichinger, Christian Erbach, Agostino Agazzari, Flaminius Comanedus, 



etc.) especially in their four-part writing, from whom a beginner can best 
learn how the harmony is to be conducted to suit the Organ Bass. For if 
a man were to turn himself to the present much ornamented (zierHche) 
vocal style and use it at the keyboard without first knowing what is con- 
formable with the rules of music, he would certainly not learn from the 
new style in question to tell good from bad, far from it, he would go in 
quite the opposite direction, since it continually requires an absolutely 
different pattern of harmony from the existing parts.' 

At this stage, information as to the harmonies going on in the 
written parts and instructions as to the composer's wishes are no longer 
necessarily the same thing. The accompanist has no obvious means of 
knowing which is which; all he sees is figures. But there is sometimes a 
certain abnormal fullness in the figuring which suggests that the com- 
poser has special wishes to impart, over and above the harmony 
already imposed by the written parts. Even these special wishes do not 
absolutely bind the performer, which would be contrary to the spirit 
of baroque music in general and of figured bass in particular; but he 
will obviously give them his respectful attention. In some of J. S. Bach's 
fullest figurings, which are really quite exceptional and stretch the 
system almost beyond its useful capacity, a strict rendering leaves little 
choice, not merely in the harmony, but in the part-writing needed to 
embody it. 

(294) A. Werckmeister, Anmerckungen, Aschersleben, 1698, sect. 70: 

'Furthermore it is not desirable just to play blindly with the singers 
and instrumentalists the discords shown in the continuo, or to double 
them : for when the singer is conveying an agreeable feeling by the written 
discord, an unthinking accompanist, if he does not go carefully, may 
ruin all the beauty by the same discord: hence the figures and the discords 
are not always written in with a view to being blindly reproduced; but 
a performer skilled in composition can see from them what is the com- 
poser's intention, and how to avoid conflicting with them with any 
matter which would be injurious to the harmony/ 

(295) Friedrich Erhard Niedt, Musicalische Handleitung, Hamburg, 
1700, Ch. V, rule 8: 

' If the singer or instrumentalist sings or plays the figures which are set 
above the continuo, it is not necessary for the Organist to play them; he 
can just play Thirds instead if that seems suitable; or, again, it is at his own 
discretion if he chooses to put in something more highly elaborated/ 

(296) C. P. E. Bach, Essay, II, Berlin, 1762, 1, 4: 

1 am no supporter of great quantities of figures. I am against everything 
which makes needless trouble and kills enjoyment. [XXV, 17:] Accom- 
panists of experience capable and courageous enough to improvise small 
corrections in a piece of music should be given praise for what they do/ 





(a) There are many occasions on which it is possible, and has an 
excellent effect, to add to the harmonies shown by the figuring. 

(297) Lorenzo Penna, Albori Musicali, Bologna, 1672, III, Ch. 14: 

'The discords can be accompanied (if so desired) by the consonances of 
the note written.' 

This is a late date to find the sounding of the chord of resolution 
with the note of dissonance unrestrictedly accepted as a desirable 
enrichment of the harmony; but see 4 (c) below for a restricted per- 
mission, and Ch. XXIX, 5, below for a wider permission under the 
special circumstances of a very full accompaniment. 

(298) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VI, p. 71 : 

'On a bass note of substantial duration, one can put in two or three 
different chords one after the other, although the text only asks for one, 
provided that one senses that these chords go with the melodic part. . . . 

'[p. 89: With two rising minor sixths the second of which is in effect a 
leading note in the bass] one can take ... on the second, the diminished 
Fifth [although not figured].' 

Ex. 168. Saint-Lambert, loc. tit., added dissonance: 

6 6 

[Ch. IX, p. 133:] 'One can sometimes add a fourth note to the chords 
indicated by the regular Rules, either in order to soften the hardness of a 
discord, or on the contrary in order to make it more piquant, so as to 
increase the pleasure of its resolving concord. What interval from the 
Bass this fourth note should comprise ... let your ear decide as occasion 
arises, and if you cannot decide, leave it out/ 

The example above is connected with a very general convention by 
which certain sixths, figured merely 6, could have the 4 added to the 
already implied 3. The most important is the major sixth on the super- 
tonic of the prevailing key (major or minor). Here the 3 is minor, and 
the chord as figured is the first inversion of the imperfect triad on the 
seventh degree of the scale. The addition of the 4 turns it into the 
second inversion of the dominant seventh. 

(299) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. Ill, 
sect. 7, p. 150: 

'In the case of the major 6th combined with a minor 3rd . . . one may 
take the 4, holding it on from the previous chord, provided it can remain 
as a stationary note into the next.' 



The first of Heinichen's two conditions (the presence of the 4 in the 
previous chord) was not indispensable; but if the 4 is not thus prepared, 
the 3 should be. 

(300) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. Ill, 
sect. 7, p. 151: 

'The same thing is allowed in certain cases with the augmented 4th s 
which, however, is then combined with the major 3rd/ 

When the sixth as well as the fourth is augmented, the result is a 
Trench Sixth' (see Ch. XXII, 9); the figuring may still be only the & 
(#6, e/c.). 

(301) C. P. E. Bach, Essay, II, Berlin, 1762, VII, i, 11 : 


'Many people figure [the 5] insufficiently with only 6 and its accidental.* 


Instead of a \ chord, a & may occasionally be taken as a 5, provided 
there is no 4 in the existing parts to conflict with it. 

(302) G. P. Telemann, Generalbass-Vbungen, Hamburg, 1733-35, 
p. 9: 

'It is, in fact, a useful attribute of the diminished 5 that one can bring it 
in on all sorts of occasions. . . . 

[p. 11 :] 'The 5 is not always figured, but it is always in keeping when 
the movement [of the bass] is upwards to a minor chord, and the composer 
has not put the 4, unfigured, in place of it; a good ear, which listens ahead, 
is needed. A 3 completes the chord/ 

But the main authorities do not describe this particular possibility, 
probably because it is very liable to go wrong in sight-reading, when 
the presence of the 4 in the existing parts may well not be foreseen. 

Telemann (op. cit., 16/and 18 h, i) also allows a 5 on the subdomi- 
nant in the absence of figures: Le. a 6 added to a perfect triad. Here, 
of course, the 5 is perfect. There are more occasions for this addition 
than for the one last mentioned, and it will often be found valuable. 

Ex. 169. G. P. Telemann, Generalbass-Ubungen, Hamburg, 1733- 
35, (unfigured) f with unfigured 6 added (on second chord): 


A dominant major triad may have a 7 added, though not shown in 
the figuring. This is always justifiable as a passing note. Struck as part 
of the chord it requires more discretion. 



(303) A. Scarlatti, Regale, Brit. Mus. MS. Add. 14244,/. 13a, Rule 1 : 

*On a I to 3 cadence in resolving on to the 3rd one strikes the 7th as 

(304) C. P. E. Bach, Essay, II, Berlin, 1762, XIII, ii, 1 : 

'When the bass, whether in a cadence or not, either rises one degree or 
leaps to the 4th above or the 5th below, the chord of the 7th, even when 
not figured, may be taken with the penultimate note, provided that the 
next note bears a triad.' 

C. P. E. Bach also gives an example in which a 6 is figured but he 
prefers to take a suspended 7 resolving on to the 6, for the sake of a 
better progression (op. cit. 9 Ch. 34, Sect. 2, Ex. a). 

(305) G. M. Telemann, Unterricht im Generalbass-Spielen, Hamburg, 
1773, III, 8, p. 104: 

'In final dominant cadences in root position one commonly strikes the 
7th after the 5th or 8th of the penultimate chord, even when not 
figured. . . . There is, however, no obligation/ 

Cases of incomplete figuring also occur, however, in which there is 
an obligation, because the figuring is left incomplete not merely to give 
the performer the option, but on the assumption that he would know 
the intention. This assumption is often most unreasonable. The in- 
adequacy of much continue figuring is eloquent testimony to the skill 
of most accompaniments in making do with the most scanty hints, or 
none at all in the case of the numerous totally unfigured basses in 
operatic and other music of the period. In a sense it increases our 
liberty; but it also adds heavily to our responsibility. Typical instances 


which are not merely incomplete but highly misleading are J for 4 or 

4, and 4 for s ; but many others are found, often quite unpredictable. 

A more comprehensible economy occurs when a figure is omitted 
because the progression implies the interval in question if it is to be 
grammatically correct. Thus a note required either for the preparation 
or for the resolution of a discord is quite commonly left unfigured on 
the assumption that the need for it will be self-evident. A special case 
of this is 9 used by itself for a chord which, besides the obvious 5 and 
3, must also include the 7 to prepare a following discord. And, of 
course, such added elements are often desirable merely to complete a 
harmony, or to enrich it, even without the further compulsion of 
securing a correct progression. The whole matter lies wide open to the 
intelligence and enterprise of the accompanist, provided that he knows 
what he is doing and can therefore keep within the boundaries of the 



(b) There are equally occasions when the accompanist may not wish 
to sound all the notes shown in the figuring. He is just as much within 
his rights in leaving out part of what is figured as he is in adding to 
what is figured. The only condition is again that just mentioned: that 
he shall know what he is doing and why he is doing it. 

Adding to the figures naturally increases the sonority; but mere 
sonority can be as effectively increased by simply doubling the intervals 
already present. The primary reason for adding further intervals is to 
increase the interest and the sophistication of the harmony. 

Subtracting from the figuring is at the same time for decreasing the 
sonority and for decreasing the harmonic sophistication. There are 
many quick and airy movements where three-part writing, or indeed 
two-part writing, is the obvious recourse, not necessarily or not only 
because the dynamic balance requires no more, but primarily because 
the character of the music required the lightest possible mood in the 
accompaniment. It is as much a matter of simplifying the harmonies as 
of thinning the texture. But, of course, beyond the point at which all 
unnecessary doubling is avoided the only further direct recourse for 
decreasing the sonority is thinning the chords themselves. Thus here 
the two things go together. 

(306) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VI, p. 71: 

'On the other hand one can avoid sounding all the intervals marked in 
the text, when one finds that the notes are too heavily loaded. 

[Ch. IX, sect. 1, p. 128:] 'This taste consists chiefly in managing the 
harmony well, so that you do not draw so much sound from the harpsi- 
chord that it stifles the melodic part entirely, or on the contrary so little 
that it does not sufficiently support it. ... For very light voices . . . you 
can even reduce the accompaniment to two parts. 

[Ch. DC, p. 133:] 'One is not always obliged to include three different 
parts in a chord, having the option of doubling [including unisons which 
on a keyboard suppress a part] some of them at pleasure, or even of 
dispensing with one of the three when that seems necessary.' 

(307) F. E. Niedt, Musicalische Handleitung, Hamburg, 1700, Ch. 
VIII, rule 8: 

'When there are rapid notes or notes with tails in the continuo, one 
must not take the Third and Fifth with each, note, nor play all the figures 
(when there are many of them over the Bass)/ 

(308) C. P. E. Bach, Essay, Berlin, II, 1762, XXXII, 6 (see Ch. 
XXIX, 4 (328), below for the continuation of this passage): 

'One of die main refinements of the accompaniment is parallel move- 
ment in thirds [tenths, etc.] with the Bass. The right hand is never obliged 
here to maintain a consistent fulness of harmony [whatever the figuring].' 



(c) It is occasionally desirable not merely to add to or subtract from 
the intervals shown by the figuring, but actually to contravene them in 
order to improve an awkward corner, such as may occur by oversight 
even in the best composers. 

Copyist's or printer's errors in the figuring are very common indeed; 
and, of course, if such an error is suspected, an intelligent performer 
will do his best not to be misled. 

(309) C. P. E. Bach, Essay, II, Berlin, 1762, Introd., 27: 

'In defective and awkward passages, where there can often be no 
proper middle part owing to the faultiness of the Bass from which such 
parts should flow, one hides the deficiencies as far as possible with a thin 
accompaniment; one is sparing with the harmony; at need, one introduces 
a [contravening] figure. 

[XXXII, 5:] 'Sensitive accompaniment means, moreover, making 
modifications to match certain liberties sometimes taken by the soloist, 
who may at his own option to some extent contradict the written notes 
when bringing in embellishments and variations. A skilled soloist will do 
this when he knows that he has a competent accompanist, so that he lets 
himself go with complete abandon to the feeling of the music. . . . The 
accompanist must modify his harmony to suit. 

[Part I, 1753, Foreword:] 'The keyboard player . . . must have at his 
command a thorough knowledge of thorough bass, which he must per- 
form with judgement, often departing from the written text/ 

Finally, there is nothing in baroque theory or practice to forbid an 
accompanist from changing the harmony shown by the figures purely 
for his own pleasure, provided that he keeps within the style and does 
not clash with the written parts. It need hardly be added that a modern 
performer has to be very sensitive in doing this, so as to avoid artistic 
incongruity whether anachronistic or otherwise. 

(309a) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VI, p. 71: 

'You can sometimes change the chords marked on the notes [i.e. 
figured], when you judge that others will suit better/ 


(a) The liberties regarded as legitimately open to the accompanist 
included altering the bass line itself if he felt that it was really unsatis- 
factory at any point. 

(310) C. P. E. Bach, Essay, II, Berlin, 1762, Introd., 27: 

'When the continue is not doubled by other instruments, and the 
nature of the piece permits it, the accompanist may make impromptu 
modifications in the bass line with a view to securing correct and smooth 
progressions of the inner parts, just as he would modify faulty figuring. 
And how often has this to be done!' 



(b) Where the bass line is satisfactory in itself, but is either not very 
suitable to a keyboard instrument as it stands, or at any rate can be 
improved from that point of view, the accompanist is entitled to make 
appropriate modifications in it. He may simplify it; he may adapt it; 
or he may elaborate it with arpeggios or divisions. 

(311) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VIII, p. 120: 

*When the measure is so compressed that the accompanist cannot 
conveniently play all the notes, he can content himself with playing the 
first note of each bar, leaving the bass viol or violoncello to play all the 

[p. 121 :] 'On the other hand if the Bass has too few notes, and drags too 
much for the liking of the accompanist, he may add other notes by way of 
pleasing figuration, provided he is sure that this will not interfere with 
the melody/ 

(312) J. D. Heinichen, General-Bass, Dresden, 1728, p. 377: 

'There is no need to be put out by a number of rapid repeated notes, 
but, so long as no fresh figuring occurs over the notes, the previous chord 
is repeated in slower notes, one or more times (according to the measure), 
and the playing of the rapid notes is left to the other accompanying bass 
instruments. . . . There are various other ways of treating such notes/ 

Ex. 170. J. D. Heinichen, loc. cit. and I, vi, 39, methods of adapting 
rapid repeated notes to the idiom of the harpsichord: 

(a) as \\iitten 

as performed I 6 $ & 

S n r-j * r-r-ii \ r-t 


(c) .as written 6 

(d) as performed 

In the above examples, Heinichen shows right-hand chords taken 
with the same rhythm as the bass. He also gives the alternative of 
quavers consistently replacing semiquavers in the left hand, while the 
right hand takes over the semiquaver movement with 'Alberti bass' 
patterns resembling those in the left hand at (d) above. C. P. E. Bach 
(Essay, II, Berlin, 1762, Ch. 29, sect. 17) gives, for slow tempos only, 
the repeated semitones of the bass played literally and accompanied by 
their lower octaves in quavers (all in the left hand). For faster passages, 



he is prepared to sacrifice the added sonority of the octave doubling; 
and when a string bass is present, he is strongly opposed to letting the 
left hand drum out repeated semiquavers, which only serve to stiffen 
the player's hand without being really effectual. As a final resort, he 
allows chords to be struck on alternate semiquavers in either hand, 
when a real noise is wanted. 
(313) C P. E. Bach, Essay, Berlin, 1753, Introd., 9 9 fn.: 

It is best to let one, three or five such notes, according to the speed and 
measure, pass unstruck, those which are struck being taken with the 
octave, or in a fortissimo, even with full chords in either hand, with a 
heavy touch and rather sustained, so that the strings may sufficiently 
vibrate, and one note be joined to the next. Or if every note must and 
shall be heard on the harpsichord, there remains the following recourse, 
namely to reproduce the movement by striking the notes [in full chords] 
with alternate hands. 

[II, 1762, XXXVI, 8:] 'When the bass has a continuous succession or at 
least a large number of demi-semiquavers or faster notes, the left hand 
may leave out one or more notes, provided a melodic bass accompani- 
ment is present. [Some authors waive this last proviso.]' 

(c) It is legitimate, and very often desirable, to introduce melodic 
'leads' to take the music neatly back to the beginning of a repeated 
passage, at the first time through, and also (with the requisite modula- 
tion) on to the next section, at the second time through. If this is not 
done, the whole piece may hang fire, and the interest flag, for lack of 
the necessary movement to keep the last bar from falling empty. These 
leads' will frequently need to be added to the bass line as well as to 
the other parts. The following examples from C. P. E. Bach, like almost 
all that he gives, were shown for economy of printing on a single stave, 
of which the top line, however, was to be taken at any required octave 
transposition (e.g. at the tenth instead of at the third). 

Ex. 171. C. P. E. Bach, Essay, II, Berlin, 1762, XXXI, 6, 'leads' 
back to the beginning of repeated sections : 


(314) C. P. E. Bach, Essay, II, Berlin, 1762, XXXVI, 13: 

'When transitional passages [in the bass] cannot be accompanied in 
thirds [or tenths, as at Ex. 171 (a) above] or some pleasing variant [as at 
Ex. 171 (b)], they should be treated as passing notes.' 

(d) Occasionally the bass has to be modified in order to accommo- 
date ornamentation in the solo part. 


(315) C. P. E. Bach, Essay, II, Berlin, 1762, XXX, 6: 

*When a composer allows his bass to continue moving through a 
closing cadence, as at [Ex. 172], regardless of the possibility of ornamen- 
tation, the accompanist [if necessary] holds on the first G and repeats it 
under the [concluding] trill, after which he begins the following bar. 
Such a case, which often happens in allegros, wants careful listening/ 

Ex. 172. C. P. E. Bach, bass needing modification if the soloist 
makes a cadenza at the point here indicated by the ^: 


O ir 

las performed according to C. P. E. Bach's instructions] 

(e) An odd case of conventional mis-notation in recitative is men- 
tioned by J. A. Scheibe. If figuring and bass appear as at Ex. 172 (a), the 
accompanist, unless somehow prevented, is almost bound to resolve 
the 2 of the \ as at Ex. 172 (b); some composers actually write it so 
yet still mean, the realisation to be as at Ex. 172 (a) (i.e. with the resolution 
transferred to the octave above, in the right hand, the bass being 
allowed to keep its rest as written at Ex. 172 (a)). The remedy suggested 
is to write the bass as at Ex. 172 (c); but it is not a very self-explanatory 
remedy, and certainly not one which shows any signs of having 
recommended itself to composers. 

Ex. 172a, b, c: a tricky case shown by J. A. Scheibe in F. W. 
Marpurg's Kritische Briefe, Berlin, for 1760-62, 87 on pp. 354-55: 

y y V V V 
i J > j 

'PPPP 3 


p 3 * i 



2 6 


4 ^ 

i- r r 



(a) Trills and long appoggiaturas have so pronounced an effect on 
the harmony that their presence has to be taken into account when 
realising the accompaniment. A few other ornaments have sufficient 
effect for it to be desirable to take them into account in some contexts. 



(b) The best accompaniment to give to a trill depends on how pro- 
longed the initial upper note (the preparation or 'appoggiatura' of the 
trill) is made. This varies partly with the style; partly with the taste of 
the soloist. 

When the preparation is very long, the accompaniment should 
normally consist of two chords, whether so figured or not. The first is 
the chord implied by the preparation (| or ^, or with two simultaneous 
trills , as the case may be). The second is the chord implied by the 
resolution on to the main note of the trill (normally 3). 

At Ex. 173 (a) below, we see a typical case of misleading figuring 
which nevertheless gives the necessary hint to an accompanist familiar 
enough with the style concerned, which is that of the French school of 
gambists. The 4 by itself would tend to suggest a 5 4 , but at least we can 
be certain that it is not a 3, which follows on the second beat with the 
# marked. We therefore know that the solo parts, which are notated to 
give a | on both the first and the second beat, are not going to be 
performed literally, but with a harmonic ornament of whose effect the 
4 is meant to warn us. And in fact both solo parts have a trill shown, 
by the sign normal to this school: a comma after the note on which the 
trill is to be made (not to be confused with Couperin's phrasing comma, 
which looks the same!). The harmonic effect of which the 4 is the 
warning will therefore be due to the long preparations of these two 
trills : and this in turn shows us that the 4 must stand not for J but for 
4. The interpretation will therefore have to be approximately as at Ex. 
173 (b). 

Ex. 173. Marin Marais, Pieces a 1 et a 2 violes, Paris, 1686, 2nd 
Suite, Sarabande, bar 8: (a) as written; (b) approximately as intended 
(my interpretation) : 

(a) [Slow] 

4 * 



At Ex. 174 (a) below we have a clearer instance of the same kind, 
in which the figuring is correct and in no way misleading; it is realised 
as it is figured, but this figuring takes into account the harmonic effect 
of an ornament in the solo part. No sign for this ornament appears; 
but that is not surprising, since the context is one in which, as both 
soloist and accompanist would have been well aware at the time, a 



cadential trill is not merely optional but obligatory. My interpretation, 
therefore, is approximately as at Ex. 174 (b). 

Ex. 174. (a) G. P. Telemann, Generalbass-Vbwgen, Hamburg, 
1733-35, No. 36, harmony figured to accommodate implied cadential 
trill; (b) my interpretation: 



J J 

ij J ^_i 



-fl I^J3J^PJ3j- =ii 

| , j 

n,,. J. 

1 I 1 

J J J J 

J J J J r > J 1 

I I 

At Ex. 175 (a) below, we have a contrary case. Here, the figuring 
shows no warning of the presence of a harmonic ornament; nor is there 
any sign for an ornament in the solo part. But the realisation of the 
accompaniment (which is original, i.e. by Telemann) shows a harmony 
which contravenes the figuring; and this harmony does in fact take into 
account the presence of a harmonic ornament on the F sharp of the 
solo part. That again is not surprising; this context, too, is one in 
which a cadential trill is not merely optional but obligatory. Even 
without Telemann's explanation, this would all have been quite clear; 
but his explanation confirms it. The passage is one which suggests a 
less prolonged preparation to the trill than Ex. 174 above; but the 
change of harmony in the accompaniment comes at the same moment: 
namely on the second half of the dominant bass (i.e. on the second D). 
My interpretation is approximately as at Ex. 175 (b). 

(316) G. P. Telemann, Generalbass-Ubungen, Hamburg, 1733-35, 
No. 5, sect. 1 : 

'The 4 intervals are taken for ornamentation, because, of course, the 
singer allows the g, that is to say the 4, to be strongly heard before the 
trill comes/ 

Ex. 175. (a) G. P. Telemann, loc. cit. 9 his own figuring contravened 
in his realisation to accommodate the implied cadential trill; (b) my 


There are two other possible ways of accommodating a trill in the 
realisation. One is by ignoring its harmonic effect entirely: e.g. by 
taking Ex. 175 (a) above not as realised, but as figured, bringing the \ 
on the first D instead of on the second. This has the result of sounding 
the chord of resolution against the note of dissonance; but if the speed 
is fairly fast, and if the note of dissonance (i.e. the initial note, or 
preparation, of the trill) is not unduly prolonged, this result is often 
surprisingly good at worst inconspicuous, at best piquant. To sound 
a dissonance against its own resolution was not normally regarded as 
good workmanship in the latter part of the baroque period; yet J. D. 
Heinichen (General-Bass, 1728, Part I, Ch. II, Sect. 54ff. 9 pp. 202 /.) 
allows some instances (9 against 8 and variants of this); and even 
C. P. E. Bach, the strictest of the great late baroque authorities, permits 
it in accommodating harmonic ornaments under certain circumstances 
see (c) in this section, below. 

The other remaining way of accommodating a trill in the realisation 
is by leaving out of the accompaniment any note or notes which would 
clash with it harmonically. This recourse is chiefly valuable in dealing 
with a slow expressive trill, when it is desired to give the greatest free- 
dom to the soloist, and where any duplication of his ornamental notes, 
however well contrived, might sound clumsy. If, for example, his trill 
is prepared on 4 and resolves on 3, a bare fifth can be played in the 
accompaniment, including, of course, the 8, and if desired even the 7, 
but leaving out both the 4 and the 3. There are passages in which 
this method is by far the best, though they will always be in the 

(c) The problem of accommodating a long appoggiatura in the 
accompaniment is very similar to that of accommodating a well-prepared 
trill; and here we have the advantage of a very careful treatment of 
the subject by C. P. E. Bach. 

(317) C. P. E. Bach, Essay, II, Berlin, 1762, XXV: 

[2:] 'Appoggiaturas can seldom be ignored in the accompaniment; they 
have usually a great share in it. 

[3:] 'Appoggiaturas retard the harmony which the bass note carries 
essentially. It is known that by the rules of good performance the appog- 
giatura is rendered loudly and its resolution softly. It is therefore doubly 
harmful to give no indication of it in the figuring; when this happens, the 
accompaniment is in most cases almost bound to be unsatisfactory. 

[5 :] 'When a bass note is figured without regard to an appoggiatura 
which goes with it, and this appoggiatura and its resolution is either 
compatible with the figures given, or the same as one of them, there is 
no necessity, even in four parts, to modify the accompaniment [e.g. 
where the appoggiatura is 4 and the figured chord 6 y in cases where this 



addition to the harmony is acceptable; or where the appoggiatura Is the 
6 and Its resolution the 5 of a ^ chord]. 

[6:] 'But when the appoggiatura is not compatible with all the intervals 
with which Its bass note is figured, because these belong to the harmony 
on to which It resolves, the appoggiatura should be played as part of the 
accompaniment, together with as many of the Intervals shown by the 
figuring as are required by the loudness of the passage and are harmoni- 
cally compatible with the appoggiatura. When the appoggiatura Is per- 
formed with great feeling, and softly, at a length solely governed by the 
wishes of the soloist, the accompanist does not play it with him, but 
sounds one or at the most two of the remaining intervals of the harmony. 
The same is often done with appoggiaturas raised chromatically by a 
semitone. Appoggiaturas in two parts at once are played as parts of the 
accompaniment, and are thus treated in three parts [in all], 

[18: Alternatively, appoggiaturas In two parts at once are omitted from 
the accompaniment.] 'One either plays the whole, or lets the right hand 

*. . . The short, [so-called] invariable appoggiatura is not included in 
the accompaniment. It does not, in fact, cause any modification of the 

Other treatments of the long appoggiatura which appear from 
C. P. E. Bach's examples include totally disregarding the appoggiatura 
and allowing it to clash with the harmony as figured; playing only part 
of the chord as the appoggiatura enters, and bringing in the remainder 
little by little, thus minimising the clash; doubling the appoggiatura, 
but resolving it in the opposite direction to the solo part; resting for 
the first half of the appoggiatura, but doubling its second half (especially 
useful in triple time the resolution forming the third beat). All short 
appoggiaturas, as he says, can be ignored in the accompaniment. The 
final recourse where no other solution can be contrived is always, of 
course, the tasto solo (bass note alone without harmonisation). 

C. P. E. Bach does not hesitate to alter the rhythm of the bass from 
even to dotted in order to avoid unpleasant or incorrect progressions 
to which the appoggiatura would otherwise give rise. Nor does he 
hesitate to add an interval to the harmony for the sake of improving it. 
(Loc. cit. 9 18.) 

As with trills, the best way of accompanying some slow and expres- 
sive appoggiaturas may well be to leave out of the accompaniment the 
intervals concerned: e.g. play a bare fifth if the appoggiatura is 4 and 
its resolution 3. But everything depends on the circumstances, and real 
subtlety is required to gain the maximum effect. 

(d) C. P. E. Bach also discusses the effect on the harmony of certain 
other ornaments, but without introducing any further principles beyond 
those discussed above. (Op. cit, Chs. XXV-XXVH.) 




(a) However adventurous our approach to the subtler problems of 
figured bass, departures from the figuring or from the given bass line 
remain the exceptions and not the rule. But there is one common 
recourse which can have a profound influence both on the harmony 
and on the melody of an accompaniment, and which yet requires no 
exceptional circumstances to justify its free use. This is the recourse of 
introducing accented passing notes and (unprepared) appoggiaturas 
into the progressions without actual displacement of the figuring. 
There is no expedient more valuable both for enriching the harmony 
and for giving flexibility to the melodic outline. 

The only limitations are the fundamental ones of keeping within the 
style and not conflicting with or in any other way detracting from the 
existing parts. It is naturally not possible to carry the use of accented 
passing notes and appoggiaturas so far when accompanying at sight as 
it is when preparing a realisation at leisure with all the existing parts 
accessible for comparison. 

(b) Unaccented passing notes have little influence on the harmony, 
but for that very reason they are all the more freely available for 
increasing the melodic interest and fluency of the accompaniment. 

(c) The following examples show passing notes (both accented and 
unaccented) and appoggiaturas. The first bar of Ex. 176 (b) shows two 
unaccented passing notes introduced as modifications of the bass line 
itself. The whole of this example was used by its author for a double 
purpose: to illustrate both passing notes and ornaments. So many 
ornaments in so short a space could not occur in normal practice. Ex. 
177 shows a wealth of accented passing notes and appoggiaturas; here, 
too, it would be an exceptional context in which so many could be used 
without interference with or distraction from the existing parts. It is 
surprising, however, what can sometimes be done in this direction; 
and Ex. 179 shows the opening of my own realisation of an actual 
movement (by Handel) in which I have carried the accented passing 
notes to an unusual pitch of elaboration, since there seemed to be a 
good opening for doing so. 

Ex. 176. J. D. Heinichen, General-Bass, Dresden, 1728, Part I, 
Ch. VI, sect. 6, passing notes added (b) to a mainly plain accompani- 
ment (a) : 

'' ri 1 i 

j JljJ 1 1 


n \ n 






J 4 N * >J 



' * 

Ex. 177. FT. Geniiniani, Art of Accompaniament^ London [1755], 
Part II, Ex. xi, realisation with numerous accented passing notes and 

' r, r ' 


j r r 

r /] j .^=p^ 

V\ ; 

i* js ' 

L\ L_ 

_ pzd 

4 3 ' 



(318) C. P. E. Bach, yj^, II ? Berlin, 1762, XXXII, 12: 

'In the use of these harmonic refinements, great care must be taken to 
avoid interfering with or covering up the solo part/ 

Ex. 178. C. P. E. Bach, loc. cit., passing notes (actual figuring 
above; theoretical figuring below; original an octave lower): 







4 3 

^76 7 65 6 1234S 

5b - 5 | 434 43- 


7 64 6 

Ex. 179. Handel, Op. I 9 No. 5, 3rd movt, my realisation (cf. 
Ex. 177 above): 



The Texture of the Part 


An important factor in a good accompaniment is varying the texture 
to suit the needs of different movements and passages. Some of the 
most valuable resources for this purpose are mentioned in the present 

Variation in the texture ranges from broad contrasts to fine nuances. 
It is an art akin to orchestration. At every bar and on every chord, a 
slightly different situation may arise with regard to the needs of balance 
and colouring. The success of an accompaniment very largely depends 
on the sensitiveness with which it is adapted to these changing needs. 
The fullness and the spacing of the chords, the degree of independent 
melodic interest, the dynamics and the registration all have their part 
to play in making an accompaniment which fits the music as the scoring 
fits the orchestral piece. The timbre and the sonority of the realisation 
have an influence only second to its actual harmony. 

The reader is referred here to the long passage from Agazzari quoted 
at (129) in Ch. IX, 3 above, and particularly to its third paragraph, 
where the importance of adapting the volume, the spacing, the colouring 
and the texture of a 'fundamental' accompaniment to the music 
accompanied is brought out with great clarity and in considerable 
detail. The date of the passage is 1607; yet it already covers all the 
primary principles of baroque accompaniment. 


Provided a melodic bass instrument is present, it occasionally makes 
an effective contrast to withdraw the harmonic continue instrument 
from a passage or movement. This may be indicated by senza cembalo 
(without the harpsichord), etc., or be done at the accompanist's dis- 

(319) C. P. E. Bach, Essay, II, Berlin, 1762, XXIII, 2: 

'In Italian works, Hghtly-composed passages often bear the indication 
senza cembalo over the bass-part by way of precaution. Entire arias may 
sometimes have this indication. 

[XXIX, 19 :] 'When several parts including the bass are played pizzicato, 
the accompanist falls silent, leaving the continue to the violoncellos and 



double-basses. If only the bass is pizzicato, however, the accompanist 
plays staccato, with his left hand [i.e. the bass line] only, unless the com- 
poser has seen fit to set figures over the notes, when the right hand 
completes the chords, also staccato/ 

A rest In the continuo line implies, normally, a rest in the accompani- 
ment, unless the rest is short and bears figures to an upper part. But 
In fugal openings where the bass has not yet entered (or later in the 
fugal movement if the bass drops out) it is open to the accompanist 
either to play the notes of the uppei entries in unison, or to remain 
silent until the bass enters. In such passages, the entries themselves may 
be shown with the appropriate clefs in the continuo part, or only the 
lowest part which has already entered, with figures supposed to show 
the higher part or parts; these methods are Interchangeable and both 
mean playing in unison with the entries, if at all. 


(a) The strict meaning of tasto solo (single touch) or tasto or the 
abbreviation /. s. is : play only the bass line till the next figures occur 
(or a change of clef). 

(320) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. V, 
Sect. 18, p. 515: 

'The words tasto solo indicate that one is to go on playing the notes 
concerned with one key or finger at a time, without other accompani- 
ment, until another part, or another clef appears. 

'In some instances, two parts only are to be seen, one above the other, 
meaning that in these instances no more is to be played than what is there 
shown [e.g. at the opening entries of fugues]/ 

(321) Lodovico Viadana, Concerti Ecclesiastici, Venice, 1602, pre- 
face, rule 5 : 

'When a Concerto opens in fugal fashion, the Organist also begins 
with a single note (con un Tasto solo), and at the entry of the several parts, 
it is open to him to accompany them at his own discretion/ 

The tasto solo is not always indicated where it is intended, or where 
it can be desirably introduced. 

(322) C. P. E. Bach, Essay, II, Berlin, 1762, XXIII, 3: 

'We use the tasto solo with excellent effect [even where it is not marked] 
in suitable passages : e.g. where bass and soloists move in thirds and sixths 
with no further parts present [or where it is desired to avoid getting in the 
way of an appoggiatura, or] with a bass carrying a melody low in pitch 
with no accompanying instruments above it [or (XXV, 17) merely where] 
it is impossible to contrive an accompaniment/ 




(b) An effect distinct from the tasto solo is the unison! or aW unisono, 
where the bass line is played on the harpsichord and also doubled at 
the octave or octaves above, and if desired, also at the octave below. 
A strict tasto solo may easily be spoiled if treated aWumsono, and the 
accompanist must use good judgment in the matter; but to some extent 
the two effects are interchangeable, and they were very naturally subject 
to confusion. It will be realised that a strict tasto solo is best taken 
without bringing in 16 ft. or 4 ft. stops unless urgently required for 
volume, but that an alYunisono is free from this restriction. 

Like the tasto solo, the alYunisono is not always indicated where it is 
intended, or where it can be desirably introduced. 

(323) C P. E. Bach, Essay, II, Berlin, 1762, Introd., 29: 

'The indications [tasto solo and a\Vunisono\ are often lacking. 

[XXII, 3:] 'Yet it is a matter for surprise that some composers do not 
always indicate their desire for an alVunisono accompaniment when 
writing out the bass. Figures are sometimes found over the bass where 
none are intended to be played. . . . [The composer deliberately plans, 
for special effect] to renounce the beauties of harmony for a time; his 
phrase is to sound in unison, and to monopolise both the thoughts and the 
fingers of all the accompanists [i.e. including any ripieno strings. But his 
pleasurable anticipations] are disturbed by the accompaniment of the 
keyboard player, who prepares and resolves the intervals figured, with 
the utmost conscientiousness and accuracy. . . . Fortunately for him the 
composer now realises that he has made a mistake in his notation of the 
bass part, and is delighted when the accompanist, disgusted with his 
inappropriate accompaniment, leaves out the harmony of his own accord, 
takes no more notice of the figures, and joins in reinforcing the passage at 
the unison, [remembering] the first principle of accompaniment: an 
accompanist shall fit to the music accompanied its appropriate harmony, 
in the appropriate strength/ 

(c) Pedal points over a held bass, even when figured, may always be 
treated tasto solo, or alFunisono. 

(324) C. P. E. Bach, Essay, II, Berlin, 1762, XXIV, 4: 

It is not easy to figure pedal points, and they are therefore generally 
handled tasto solo. Those who do figure them must put up with their 
being performed tasto solo just the same [because of the extreme difficulty, 
if not impossibility, of reading the unusual figurings required]/ 

The difficulty varies, of course, with the complexity of the harmony 
composed above the pedal point. When this harmony changes fre- 
quently and rapidly, a tasto solo or alPunisono makes the best effect, 
quite apart from the difficulty of doing otherwise. When, however, the 
harmony is simple and changes not too frequently or rapidly, it may 
often be desirable to accompany with full chords, whether these are 



shown by figuring or not. It may often happen that the parts above the 
pedal point, though melodically in rapid motion, are mainly executing 
broken chords, and that the chords change infrequently enough for the 
accompaniment to keep up with them quite easily if it takes the chords 
plain and without figuration. Indeed, the chords themselves may be 
mere alterations of dominant and tonic harmony. But whatever they 
are, if they move slowly enough to sound well when supported by the 
harpsichord in plain harmony, this support may be given at the 
accompanist's discretion. 


(a) The unharmonised accompaniment is naturally the rare excep- 
tion. In the normal circumstances requiring harmony, the question 
arises as to how thick or thin this harmony should be. 

There is a range of choice here from two parts to as many as the 
fingers can manage. This entire range should be exploited to the full, 
not merely as a means of contrasting one movement with another, but 
also for internal contrasts within movements. 

(b) There is evidence that some organists, at any rate, of the seven- 
teenth century preferred a three-part accompaniment when the ensemble 
was small. 

(325) Michael Praetorius, Syntagma, III, Wolfenbiittel, 1619, Ch. 
VI, sect, on Organ, App.: 

'When there are few voices singing, few keys [notes] should be touched, 
as eg e l , d af f c c l e l etc., so that the voices can be heard clearly and dis- 
tinctly above the organ; but when more voices begin to sing, more keys 
and fuller harmony should be used. 5 

On the other hand Praetorius' own example (loc. tit.) is in four parts, 
and we should be wrong to infer a general early baroque preference for 
three-part writing such as Heinichen seems to imply in his reference 
(1728) to the 'old musicians', at (354) in 8 (a) below. 

(c) There are movements so light that the best harmonisation is in 
two-parts, or mainly so ; others for which a mainly three-part accom- 
paniment is the most satisfactory. 

(326) C. P. E. Bach, Essay, II, Berlin, 1762, Introd., 26: 

'Accompaniments in three or fewer parts are used in lighdy-composed 
music where the style, the execution or the feeling of the piece suggests a 
sparing use of harmony/ 

Another possible (but by no means necessary) reason for thinning 
the accompaniment is a bass which goes high. 

(327) Johann Staden, Kurzer Bericht, in his Kirchen Music, II, 
Nuremberg, 1626: 



'When the clef [of the continue] Is Soprano or Alto, few parts are 
required [in the accompaniment].' 

We shall see at (g) below that mere decrease of volume is another 
reason for thinning the accompaniment, either momentarily or through- 
out, just as increase of volume may be a reason for thickening it. 

A particular form of two-part or three-part accompaniment is that 
in which a running bass is accompanied mainly in tenths, etc. 

(328) C. P. E. Bach, Essay, II, Berlin, 1762, XXXII, 6: 

4 One of the main refinements of accompaniment is parallel movement 
in thirds [tenths, etc.} with the bass. The right hand is never obliged here 
to maintain a consistent fullness of harmony. Four-part writing is seldom 
maintained, excepting with slow notes, since thirds cannot be clearly 
brought out in four-part writing at a rapid speed. Three-part writing, and 
indeed in most cases two-part writing, is more satisfactory.' 

[7:] 'Plain chords should be played, and movement in thirds abandoned, 
when the solo part already has these thirds or other intervals moving in 
the rhythm of the bass. . . . 

[8:] 'Sometimes sixths may be mixed in with the movement in thirds 
. . . [and if there are to be three parts, the whole may be bound together 
by the addition of] sustained notes/ 

(d) Other considerations being equal, a mainly four-part accompani- 
ment is the normal standard. 
(328a) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. Ill, p. 19: 

'You play the bass with the left hand, and to each note of the bass 
which you touch, you add three others in the right hand, thus making a 
chord on each note [this is a general statement and is later subjected to 
many qualifications]/ 

(329) Joachim Quantz, Essay, Berlin, 1752, XVII, vi, 4: 

'The standard rule for thorough-bass is to play regularly in four parts, 
though for really good accompaniment it is often better not to follow 
this too consistently, but instead to leave out some parts or even double 
the bass at the octave above in the right hand [Le. play alVunisono\! 

(330) C P. E. Bach, Essay, II, Berlin, 1762, Introd., 24: 

'Consistent four-part writing, or more, is used in music which is 
thickly composed, or is in the learned style with counterpoint, imitations 
and so forth, or particularly in music composed for the music's sake with 
little scope for taste [of the galant kind]. 

[25:] *To keep the spacing of the parts satisfactory it is better to bring 
two parts to a unison than to insist on four awkwardly distinct parts at 
the cost of pointless leaps and clumsy progressions. 

'[37: We should when necessary have recourse to] first, playing a 



middle part [as well as the bass] in the left hand; second, bringing two 
parts to a unison; third, adding a fifth part momentarily to avoid con- 
secutive fifths while at the same time changing the spacing; fourth, 
changing the spacing by repeating the same chord [in a higher position] 
over a single bass note, with a view to getting back into a higher position 
if the accompaniment has got too low [the same may be done in reverse]. 

[Ch. II, i, 37:] 'If the hands get too close together, or the right hand 
gets too low down, the same chord may be repeated in a higher position 
over the same [bass] note, if this note is not too quick; if there is no time 
for that, an additional part may be taken in at the top, and the lowest one 
[not counting the bass dropped]. This recourse is to be used (1) only in 
case of need, since I believe that as a normal procedure four regular parts 
should be maintained without unnecessarily adding to them; (2) only 
with consonances, since discords impose their own limitations on the 
accompaniment. ' 

If it is desired to retain four-part writing rather than momentarily 
increase the number of parts to five, an almost identical effect may be 
obtained by supposing two of the parts to cross. None of the stricter 
authorities were prepared to admit this device as a means of avoiding 
forbidden consecutives which it leaves audible; but where no such 
forbidden consecutives are involved, it can be used to gain a higher 
position than could otherwise be taken without incorrectly resolving a 

Ex. 180. Momentarily taking in a fifth part to gain a higher 

Ex. 181. Crossing of parts to gain a higher register: 

(331) Johann Philipp Kirnberger, Gmndsatze des Generalbasses, 
Berlin [1781], Sect, after 172: 

'With a succession of ninths, instead of taking in a fifth part, we can 
also allow the parts to cross without interfering with the correctness of 
the part-writing/ 



(e) The regularity of the part-writing is not disturbed by adding 
octaves in the bass for greater prominence and sonority. 

(332) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 7: 

'Merely doubling the bass [at the octave] has also a penetrating effect, 
and is essential where the notes are not very quick and can readily be 
executed [in this way], yet have a definitely melodic character of some 
considerable duration. When there is a [bass] entry in a fugue, or any 
imitation which needs bringing out, this doubling of the bass is extremely 
effective. But if an entry or any other phrase needing special prominence 
includes ornamental figuration which cannot readily be played in octaves 
by one hand, the main notes at any rate should be doubled and the 
remainder played singly.' 

(333) G. M. Telemann, Unterricht im Generalbass-Spielen, Hamburg, 
1773, Introd., 8: 

If the bass lies too low to double at the octave below, we must either 
not double at all, or double at the octave above.' 

It is also quite often useful to double intermittently at the octave, 
sometimes adding the fifth in the left hand, and sometimes alternating 
octaves and fifths. 

(f) When great sonority or great weight is required in the accom- 
paniment, both hands may play as many notes as their fingers can 
accommodate. This is known as the filled-in accompaniment, and is 
necessarily exempt from most of the rules of correct part- writing which 
apply to the normal accompaniment. It may be used for entire move- 
ments of a suitable character; for certain passages in a movement 
otherwise normally accompanied; or for single chords to which it is 
desired to give additional volume and emphasis. The two chief principles 
on which a filled-in accompaniment is based are : keep a good progres- 
sion between the topmost part and the bass (Le. between the two outer 
extremes) ; and keep the thumbs close together (i.e. leave as little space 
as possible between the chords taken by the two hands, so that there 
is no room for prohibited consecutives to become conspicuous in the 
middle of the harmony). 

(g) It must be realised that the subtraction or addition of parts is 
one of the chief means by which the harpsichord, whose control of 
volume by touch is very slight, can make dynamic variations, either 
momentary or continued: see (306) in Ch. XXVIII, 2 above. The other 
chief means is by the registration, where the instrument possesses stops 
which can be brought on and off. This latter means has also, of course, 
a crucial effect on the colouring of the tone. On the simplest harpsi- 
chords it is not available. 



(334) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 5: 

'Of all instruments used for continue playing the harpsichord with one 
manual is the hardest to manage for forte and piano. To meet the defi- 
ciencies of the instrument in this matter, the number of parts must be 
increased or decreased. Yet care must be taken to include all necessary 
notes and to avoid any incorrect doublings. [A staccato touch is a bad 
solution;] it is better to decrease the volume [by striking fewer chords] 
in the right hand over passing notes [in the bass]. 

'[7: On two-manual harpsichords] the fortissimo and the forte are 
taken on the stronger manual. [For the fortissimo, the filled-in accom- 
paniment can be used: i.e.] doubling in the left hand all the notes of 
consonant intervals, and the consonant intervals only of dissonant chords, 
so far as the playing of the bass permits. ... In the mezzo forte the left 
hand may play the bass, undoubled, on the stronger manual, while the 
right hand accompanies on the softer. In the piano, both hands play on 
the softer manual. The pianissimo is taken on this latter manual, but with 
a decreased number of parts. 

[10:] 'Modulations are brought out by reinforcing the accompaniment. 
Thus in a fortissimo, for example, both hands take a filled-in chord 
rapidly spread upwards, the bass and its octave, together with the notes 
in the right hand, being held on. ... Distinguish the most important 
notes by reinforcing their chords, perhaps in both hands. 

[13:] 'When the solo part has a long sustained note which by the con- 
ventions of good style should start pianissimo, swell gradually to fortis- 
simo, and fall back in the same way to pianissimo, the accompanist must 
follow it with the greatest accuracy. Every recourse open to him must be 
used to make the forte and the piano. His crescendo and diminuendo 
must correspond with the solo part, neither more nor less/ 


(a) The standard of grammatical correctness, in so far as this applies 
to accompaniment, should be about as strict as the composer's own 
standard, so that the two shall match in style. A composer of somewhat 
classical temperament such as Corelli needs to be accompanied rather 
more correctly than a composer such as Purcell whose own revolu- 
tionary boldness not merely invites but demands a comparable boldness 
in the accompaniment. 

Within the boundaries of the style in question, the degree of strictness 
is a matter of individual preference. The baroque authorities themselves 
show a remarkable variability, and it must be remembered that the 
rules themselves are only generalisations about what experience shows 
will sound well in practice. 

(335) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VIII ('On 
the Liberties which may be taken when accompanying'), sect. 13, p. 126: 



'Since Music is made only for the ear, a fault which does not offend it 
is not a fault.' 

(b) The rule forbidding consecutive octaves and fifths is funda- 
mentally valid in accompaniment, but with important exceptions. 

So long as the ideal of accompaniment in polyphonic music remained 
a literal reduplication of the parts, consecutive fifths were inevitable 
on the keyboard wherever these parts produced them correctly by 
crossing; for crossing of this kind has no audible reality on keyboard 

(336) Lodovico Viadana, Concern Ecclesiastic!, Venice, 1602, rule 

'The organ accompaniment is never under any obligation to avoid two 
fifths or two octaves [consecutively], but the parts sung by the voices are/ 

As the ideal of reduplicating the existing parts gave place to the 
typical concept of figured bass, which is a harmonic texture conforming 
to but not reduplicating the existing parts, we find a somewhat increas- 
ing tendency to strictness here. 

(337) Lorenzo Penna, LiPrimi Albori, Bologna, 1672, Book III, Ch. 
I, rule 10: 

'Two consecutive octaves, and two fifths, together with their com- 
pounds, are to be avoided whether occurring by step or by leap ; which 
rule is mainly applicable to the outside parts/ 

(338) Matthew Locke, Melothesia, London, 1673, rule 1 : 

'But (for prevention of glutting or offending the Ear) never ascend or 
descend with two Fifts, or two Eights together between the Treble and 

Yet composers of Locke's period, and to some extent into the 
eighteenth century as well, regarded the retardation of one of the parts 
as quite sufficient to rectify consecutive fifths or octaves: for example 
in preparing a ninth (even in the top part) on the octave of the bass, 
and then resolving it also on the octave of the bass. 

Ex. 182. Consecutive octaves justified by retardation: 




(339) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VIII, p. 125: 

'[Two consecutive octaves or fifths are harmless] when you are accom- 
panying a big musical ensemble, where the noise of the other instru- 
ments [covers them]. ... But when you are accompanying a single 



melody, you cannot attach too much importance to correctness, especially 
If you are the sole accompaniment; for then everything shows . . . 
[two consecutive fifths are correct] provided that the first is perfect, the 
second diminished or augmented [or vice versa if not to the bass]. 

'Finally, It would only be taking a slight license for one part to make 
even a perfect [as opposed to a diminished] fifth twice in succession witli 
another part. 

1 know that the greatest strictness would not have It; but since this 
fault (if It is one) does not show at all, I hold that one can make It 
unhesitatingly !* 

(340) F. W. Marpurg, Handbuch bey dem Generalbasse, I, Berlin, 
1755, 2nd ed. 1762, Sect. II, ixa, p. 99:' 

'[Progression from diminished to perfect fifths Is allowed in three or 
more parts] but only in the middle parts, or between an outside and a 
middle part. On the other hand, whatever the scoring, and between 
whatever parts, one may always progress from a perfect to an imperfect 
[i.e. diminished] or an augmented fifth/ 

(341) C. P. E. Bach, Essay, II, Berlin, 1762, Introd., 36: 

'While awkward progressions may be tolerated, Incorrect ones can- 
not . . . forbidden [consecutive] fifths and octaves must be strictly 

[38:] 'Unavoidably awkward progressions, hidden fifths and octaves, 
and certain permitted consecutive fifths with the bass must be kept in the 
middle parts. The top part must always flow and keep a correct relation- 
ship with the bass. 

[II, i, 22:] 'A perfect fifth may fall to a diminished fifth between any two 
parts. But a diminished fifth may fall to a perfect fifth only in case of 
necessity, and not, if it can be avoided, between outside parts. 

[II, i, 23:] 'In ascending . . . both progressions [i.e. perfect rising to 
diminished fifth or diminished rising to perfect fifth] are correct only 
between middle parts/ 

It is to be observed that C. P. E. Bach was one of the strictest of all 
baroque authorities in matters of musical grammar, and was unusually 
sensitive to consecutives occurring even in the middle parts and under 
circumstances where almost any other authority would certainly have 
excused them. A more typical baroque attitude towards accompaniment 
was to excuse, without actually encouraging, any consecutives which 
are not conspicuous to the ear in practice. This attitude produces very 
satisfactory results, and I have no hesitation in recommending it, as 
did Saint-Lambert at (339) above, or Marpurg at (340) and again very 
sensibly as follows. 

(342) F. W. Marpurg, Handbuch, I, Berlin, 1755, 2nd ed. 1762, II, ii: 



1 would [merely] say that a disposition of the parts which is free from 
all those [consecutive] octaves and fifths which are excusable is preferable 
to one in which such excusable octaves and fifths occur/ 

To pretend that two of the parts have crossed was one of the most 
accommodating excuses, but we have already seen how little reality this 
pretence usually possesses, and it was only accepted by the laxer 
authorities (e.g. Johann Mattheson, Organisten-Probe, Hamburg, 1719), 
and even by these it was not held to cover consecutives between the two 
outside parts. Others (e.g. F. W. Marpurg, Handbuch, I, Berlin, 
1755, II, ii) admitted it only in the case of obviously acceptable con- 

In harmony of four or more parts, or of fewer parts at a rapid speed, 
consecutive fifths or (but less readily) octaves in contrary motion even 
between outside parts were tolerated (e.g. by F. W. Marpurg, loc. eft., 
Figs. 27 to 29). 

Consecutive octaves or unisons between the accompaniment and a 
solo part were regarded as legitimate doubling, and in no sense pro- 
hibited, although for reasons of taste, not connected with musical 
grammar, such doubling was not favoured as a continued recourse after 
the very early baroque period. It should, indeed, be used very sparingly, 
and either from positive necessity, or as a special effect of scoring. 

Consecutive fifths between the accompaniment and a solo part, on 
the other hand, are apt to sound unpleasant, but may be impossible to 
avoid when sight-reading a continuo part printed without the solo part 
or parts above it, for the simple reason that there is nothing to forewarn 
the accompanist that they are about to occur. 

(343) G. Ph. Telemann, Generalbass-Ubungen, Hamburg, 1733-35, 
No. 23: 

'For if the player were only guided by the figures, not having the score, 
he would play [such consecutive fifths], without being blameworthy.' 

This last consideration is of the utmost general interest. Very many 
problems in the realisation (on paper) of a figured bass are solved in 
the best and most natural way simply by asking oneself the practical 
question: what would or could an accompanist make of the passage if 
he were reading it at sight? A great many clashes between accompani- 
ment and solo parts which look most alarming on paper resolve them- 
selves without even being noticed in performance. An editor may wrack 
his brains for a correct solution and produce a most tortuous and un- 
convincing result, whereas if he merely sets down what the figures 
allow him to know at sight in the absence of the score, he will find a 
straightforward progression which perfectly meets the case. It is a great 



mistake in figured bass accompaniment, as in other matters, to try to 
be too clever. 

(c) One of the best and oldest hints for avoiding unacceptable con- 
secutive fifths and octaves is to make great use of contrary motion. 
This hint appears in most treatises on figured bass throughout the 
baroque period. 

(344) Matthew Locke, Melothesia, London, 1673, rule 10: 

'For the prevention of successive Fifts and Eights in the Extream 

Parts ... the certainest way for the Beginner, is, to move his Hands by 
contraries : That is, when one Hand ascends, let the other descend/ 

(344a) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. V, p. 65: 

'The hands should [with various qualifications] always move in opposite 
directions [to prevent unwanted consecutives]. 

[p. 66:] 'When the bass makes a wide interval [a fourth or more] in the 
progression of the notes, the two hands can make similar movement 
[since this does not conduce to unwanted consecutives].' 

(d) Other things being equal, the more correctly the discords are 
prepared and resolved, the more satisfactory the accompaniment will 
sound. The chief gain is in the smoothness of the effect. Proper prepar- 
ation and resolution give a sense of logical inevitability and stability 
as nothing else can do. These qualities are essential for music in the 
learned style. For more relaxed and informal music they are by no 
means so necessary, and it was recognised throughout the baroque 
period that many more liberties could be taken outside the learned 
style. Towards the end of the baroque period this distinction crystallised 
in the differentiation of the learned or 'old' style from the galant or 
'new' style. But, of course, not all earlier baroque music had been 
learned music, and the distinction itself was not a novelty. 

In the less strict forms of music, the preparation of a discord may be 
either shortened (contrary to the strict rule by which the preparation 
must be at least as long as the discord) or dispensed with. Wholly 
unprepared discords are extremely common in many baroque styles of 
music, and the accompanist can allow himself great freedom in this 
respect wherever the musical context suggests doing so. 

The resolution of a discord may be either transferred to another part 
(not necessarily in the same octave), or suppressed in such a way that 
the mind can in some sense supply it without its being heard. In the 
former case the resolution is still present, although not in its normal 
situation. In the latter case the resolution is not so much dispensed 
with as ellided. A substitute 'resolution' by an upwards instead of a 
downwards progression is also not uncommon. The freedom to dispense 



entirely with the resolution of a discord in baroque music is noticeably 
less than the freedom to dispense with its preparation. 

Ornamental resolutions are, of course, perfectly correct, within limits 
for the learned style, almost without limits for the free style. 

The following excellent advice to accompanists can be confidently 

(345) G. M, Telemann, Unterricht im Generalbass-Spielen, Hamburg, 
1773, I, ii, 9: 

'All the rules just given with regard to preparation and resolution are 
to be observed with care so long as they can be observed. . . . For since, 
in compliance with the figures, discords are often not prepared, and often 
not resolved . . . and since, moreover, discords which ought strictly to 
resolve downwards have sometimes, in compliance with the figures, to 
resolve upwards, and the other way about, and so on and so forth ... I 
say: prepare so long as you can prepare, and again, resolve so long as you 
can resolve . . . and for the rest leave the responsibility with the author 
of the figures and the composer.' 

(e) Throughout the baroque treatises, rules are found forbidding the 
doubling of accidentally sharpened major thirds, or notes of the bass 
which are accidentally sharpened. These rules, which are often unclear 
or unsatisfactorily formulated, have as their foundation the undesir- 
ability of doubling the leading note of the prevailing key, whether 
diatonic or accidental, and in whatever part (including the bass) it may 
appear. It is the nature of a leading note to lead: i.e. to progress by a 
semitone upwards to its (temporary or permanent) tonic. If it is 
doubled, one of the parts to which it is given must progress wrongly if 
consecutive octaves are to be avoided. The only escape from this 
difficulty is by way of the licence permitting a leading note in an inner 
part to drop a major third to the fifth of the key. But in a normal four- 
part accompaniment it would sound extremely crude to use this license 
at the same time as the normal progression upwards by a semitone. 
The rule forbidding this doubling (which in any case sounds ugly in 
itself) is one which should be respected to the full in all ordinary 

The following statement of the rule, though late in date, is quoted 
because it is put clearly and accurately, and it may be taken as generally 
reliable for the baroque period as a whole. Some of the earlier formula- 
tions are unnecessarily wide, and if taken literally prevent some per- 
fectly harmless doublings. The standard implied by the following will 
be found adequate in any typical baroque context. 

(346) Johann Philipp Kirnberger, Grundsatze des Generalbasses, 
Berlin [1781], III, 154: 

'Theorists agree in forbidding the doubling of the major third; but 



this rule is only applicable to the dominant chord, in which this third Is 
the leading note to the following semitone above; hence It also must not 
be doubled when It comes by reason of an accidental chromatic to the 
prevailing key, since It then becomes the leading note of the note next 
following. In the triad of the main note or tonic, and the triad of the 
subdominant, it can be doubled unhesitatingly, unless special circum- 
stances point to the contrary/ 

There have always been some authorities who dislike the mere 
doubling of the major third In a triad, but while this is certainly not 
quite satisfactory to the ear in many contexts, It hardly amounts to a 
fault of grammar. It would not usually be one's first choice, but it 
need by no means be excluded In every instance. 

(f ) It has commonly been regarded as a fault of grammar to sound 
the note of resolution at the same time as the discord about to resolve 
on to it: or in other words, to anticipate the resolution of a discord in 
the harmony accompanying it. Yet this procedure was not merely 
sanctioned, but encouraged and practised, by some authorities and 
composers at all stages in the baroque period, and Indeed subsequently. 

(347) Lorenzo Penna, Li Primi Albori, Bologna, 1672, Ch. 14, end: 

'To complete the accompaniment of the [bass] notes on which stand 
discords to be resolved on to the concords, let the student know that it is 
a general rule that one can or rather should add to the discords the thirds, 
fifths, octaves and their compounds of the [bass] note written, making 
them all sound at the same time [more particularly if the music is in many 
parts], for even the composer writes them so.' 

In practice, it is only in a full accompaniment that a discord can be 
accompanied by the entire concord on to which it is about to resolve. 
Half a century after the above quotation, we find this recognised in 
theory also : see (349) in 6 (c) below. It became unusual, moreover, to 
make such full clashes, except in the case of some suspensions. It was 
usual at most to sound the note rather than the entire chord of resolution 
against the discord. 


(a) The degree to which grammatical correctness may be dispensed 
with in the filled-in accompaniment was viewed somewhat differently 
by different authorities, but from the nature of the case it is obvious 
that very little except the relationship of the outside parts can really be 
kept at all ordinarily strict. With so many notes going on, and no 
possibility of genuinely crossing the parts to avoid faulty progressions, 
part-writing in the ordinary sense is not in question, and we are dealing 
with something much more like the doublings in orchestration. 



(b) The two main rules are: 

First: keep a good progression between the outside parts. This is 
very important and should never be disregarded. 

Second: keep the thumbs close together so as to avoid an awkward 
space between the hands in which conspicuous forbidden consecutives 
may occur. This is highly practical, and should only be disregarded for 
a definite reason. But for special effect the hands may occasionally be 
taken wide apart. 

When this is done, the sonority of the left hand wants careful atten- 
tion. If it is taken down to a low register, it should not under any 
normal circumstances be given full chords, but only octaves or octaves 
and fifths. This is simply because of the acoustic fact that close scoring 
at a low pitch makes a thick effect. This thickness is less marked on a 
harpsichord (with its clear tone, due to the high proportion of its tonal 
energy going into its upper harmonics), than on the modern piano 
(with its massive lower harmonics) or on the organ (with its sustained 
sound). Nevertheless, even on the harpsichord, a third is too narrow 
an interval to sound at all natural below a certain pitch, and full chords 
sound still more unnatural. The cases in which such an excess of 
bottom-heavy thickness could be effective must be rare indeed, if they 
exist at all. 

Full chords at the top of the harpsichord, on the other hand, can 
sound most effective; but the effect is still a special effect, particularly 
if the left hand is some distance away instead of being close up to give 
support. Like most special effects, this one has its uses; but they will 
not be at all common. 

A third rule, then, is that the fiUed-in accompaniment will normally 
sound best when the two hands are not only close together but some- 
where within the middle register of the harpsichord. 

A fourth rule concerns the treatment of discords. Most authorities 
allowed the dissonant notes as well as the concordant notes to be 
doubled in both hands alike; some allowed only consonant notes to be 
doubled in the left hand, and dissonant notes to be doubled in the right 
hand alone, but there is certainly no need for a modern accompanist 
to observe this restriction unless he so desires. 

Any doubling of dissonant notes makes it impossible to resolve them 
correctly in each part alike, except by allowing the parts to move in 
consecutive octaves. This was, in fact, very generally permitted: some- 
times only in the left hand, the discords in the right hand being expected 
to behave correctly; sometimes in either hand. 

The alternative to allowing the doubled dissonant parts to progress 
correctly, but in consecutive octaves with one another, is to allow one 
of them to progress correctly, but the other freely. This, also, was very 



generally permitted: sometimes in both hands; sometimes in the right 
hand alone. The degree of freedom should not, however, be in excess 
of the need: e.g. there is no need for the doubled dissonance to move 
by leap if it can, with a much nearer approach to correctness, move 
upwards (instead of downwards) by step, etc. 

For general modern purposes, the rule concerning the treatment of 
discords in a filled-in accompaniment may perhaps be stated in the 
following form : where discords are doubled, one part (normally the 
higher) should observe much the same standard of correct preparation 
and resolution as in an ordinary four-part accompaniment; the other 
part can be taken in either hand, and has the choice of moving in 
consecutive octaves or quite freely. 

(c) Of the following authorities, the first, a pupil of J. S. Bach (whom 
he here describes), gives a remarkable confirmation of the length to 
which the filled-in accompaniment might be carried; and it is, indeed, 
one of the most important resources of a well-equipped accompanist. 
The second was about average for strictness; the third, stricter than 
average. We have to allow for all the usual baroque variability; but 
sonority rather than correctness is the primary consideration here. 

(348) Johann Christian Kittel, Der angehende praktische Organist, 
Part III, Erfurt, 1808, p. 33: 

'When Seb. Bach performed a piece of Church music, one of his most 
able pupils had always to accompany at the harpsichord. One can well 
imagine that one dare not put oneself forward with a sparse realisation of 
the continue, in any case. Nevertheless one had always to hold oneself 
prepared frequently to find Bach's hands and fingers suddenly mixed 
with the player's hands and fingers, and, without further interfering 
with these, decking out the accompaniment with masses of harmony 
still more imposing than the unsuspected near proximity of the stern 

(349) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. II, 
Sect 30, p. 131: 

'[Experienced accompanists] commonly, and particularly on such 
instruments as harpsichords, try to reinforce the harmony yet further, and 
to accompany in as many parts with the left hand as with the right, 
according to the lie of the two hands, in an accompaniment of from 6 or 
7 to 8 parts. 

[En. :] 'On such instruments as harpsichords, the fuller the accompani- 
ment in either hand, the more satisfying [in suitable contexts] the result. 
With organs, however, especially in lightly-composed music, it is as 
well not to get too fascinated by an over-full accompaniment in the left 
hand, since the constant growling of so many low notes is distressing to 
the ear. 



[32:] 'One has only to take care to devise the outside [upper] part of the 
right hand so skilfully that it moves with the bass without any fifths, 
octaves or other incorrect progressions. 

[35:] 'The great art of a filled-in accompaniment consists in managing 
the two outside parts. 

[38:] 'The closer the hands are kept to each other, so that no undue 
distance or empty space lies between the middle [adjacent] parts of the 
right and left hands, the more satisfying will be the resulting accompani- 

[Ch. Ill, Sect. 54, p. 202:] 'All discords in the right hand should pro- 
gress correctly [but those in the left hand may either double the right 
hand's progressions, or move freely]. 

[54:] 'In certain cases, the left hand may anticipate the resolutions of the 
discords taken in the right hand. 

[55 :] It is well known that with the 9 taken in full harmony among 
other middle parts, one can also strike the octave ... on which the 9 
must resolve. 

[56:] 'But if the ear can accept the anticipation of the resolution of the 
nine, it must necessarily also be able to accept all other anticipations of 
resolutions which are produced merely by the inversion of this 9 in full 

[61 :] 'With double suspensions, one can also introduce double anticipa- 
tions of the resolutions, but infrequently and with discretion, so that the 
cumulative dissonances shall not be unbearable to the ear/ 

(350) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 7: 

'[The filled-in accompaniment consists in] doubling in the left hand all 
the notes of consonant intervals, and the consonant intervals only of 
dissonant chords, so far as the playing of the bass permits. This doubling 
must not be done low down, but close to the right hand, so that the chords 
of both hands adjoin, and that no space may be left between them, since 
otherwise the growling low notes will make a hateful confusion/ 


(a) During the seventeenth century, no reliable generalisation can be 
inferred as to the spacing of accompaniments; but in the eighteenth 
we find a distinct tendency to regard as standard the playing of the bass 
alone in the left hand, and the remaining parts in the right hand. 

(351) Friedrich Erhard Niedt, Musicalische Handleitung, Part I, 
Hamburg, 1700, Ch. VI, rule 1 : 

'The written-out continue bass is played with the left hand alone, the 
remaining parts (whether shown by figuring or not) with the right 



(352) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. II, 
Sect 29, p. 130: 

'[Four part accompaniment came to be] divided unequally between the 
hands, that is to say three parts to the right hand, and only the bass part to 
the left/ 

[30:] 'This kind is nowadays the most general and fundamental accom- 
paniment, which every beginner is taught.* 

(b) At the same period we find the more open scoring, by which the 
parts are divided fairly evenly between the two hands, regarded as a 
distinct kind of accompaniment under the name of the divided or 
extended accompaniment. This is the converse of the full or filled-in 
accompanied just described, and like that was no novelty at the time 
when it began to attract theoretical attention. 

An extended spacing may be a momentary necessity to avoid awk- 
ward part-writing or forbidden consecutives; but primarily it is an 
effect of scoring used for its own sake, either in passing, or throughout 
a passage or a movement. As a resource of accompaniment, it is as 
valuable as its opposite, the filled-in accompaniment; but it requires no 
special consideration, being subject to the ordinary rales of four-part 

(353) C. P. E. Bach, Essay, II, Berlin, 1762, XXXII, 10: 

'An extended accompaniment ... is often one of the most pleasing 
refinements. In previous chapters we have already shown the occasional 
necessity for using it [for the sake of the part-writing]. Apart from such 
necessity, it is a familiar fact how very pleasing extended harmony can 
sometimes be in contrast to close harmony.' 


(a) It was generally agreed that the accompaniment should not 
persistently reach a very high or low register. 

The following suggests a reason for not going high which may well 
have influenced those accompanists who (like many organists of the 
seventeenth century) preferred a thin accompaniment when the en- 
semble was small (see 4 (b) above; but, of course, full accompaniments 
were also frequent at the same period, and Heimchen's information is 
to that extent misleading, no doubt because he is not giving it at first 

(354) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. VI, 
Sect. 28, note h, p. 548: 

* When the old musicians gave the rule that the right hand should not 
readily go above c 11 [in the treble stave] and on no account above e 11 [top 
space of treble stave], they were trying in this way to avoid in their 




accompaniments, which were then very thin, and most often only in 
three parts, too large a distance, or empty space between the hands, and 
in this they were well advised/ 

The reason for not carrying the harmony too low, as we have seen, 
is primarily acoustic: the narrower intervals, such as thirds, beat more 
roughly at low pitches, and therefore sound less pleasing to the ear, so 
that on the whole a scoring which keeps these narrower intervals 
moderately high in pitch is more satisfactory than the opposite. It is 
therefore a natural tendency, other things being equal, to give most of 
the harmony to the right hand, though low dispositions are also found. 
They were commoner in the seventeenth century, before this tendency 
to restrict the left hand to the bass alone became something of a 
convention see 7 (a) above but they occur at all periods. 

For most baroque music, the following general limits for the right 
hand are characteristic enough, though possibly a little wide for some 
of the earlier schools. 

Friedrich Erhard Niedt (Musicalische Handleitung, Part I, Hamburg, 
1700, Ch. VI, rule 2): between c 1 (middle C) and c 11 (the octave above 
middle C) in the main, and not below a or g (below the treble stave) 
or above e il or/ 11 (at the top of the treble stave) at the outside. 

Michel de Saint-Lambert (Traite, Paris, 1707, Ch. V): up to e 11 or the 
highest/ 11 (i.e. the top space or top line of the treble clef: but higher 
when the bass is high, i.e. becomes a bassetto: see (e) below). 

F. W. Marpurg (Kunst das Clavier zu spielen, Berlin, 1750, II, ii, 24, 
p. 36) : not usually below e (in the bass stave) or above g 11 (note above 
the treble stave). 

(b) It was also understood that these limits, particularly the upper 
one, might be exceeded for special effect, or for special reasons. 

(355) C. P. E. Bach, Essay, II, Berlin, 1762, II, i, 24: 

*The right hand should not ordinarily be taken above/ 11 , unless the bass 
goes very high, or a special effect is to be made in the higher registers. 

[25 :] 'The right hand should not be taken below the middle of the 
tenor octave [c 1 to fl], except under the opposite conditions to those 
mentioned in the previous paragraph [i.e. when accompanying a solo 
which itself is low in pitch, as on the violoncello, etc.]. 9 

(c) The general preference was to keep the top part of the accom- 
paniment below the solo part (or below the highest solo part in trio 
sonatas, etc.). 

(356) Joachim Quanta, Essay, Berlin, 1752, XVII, vi, 21 : 

It has long been a weE-established rule, that in playing continue, the 
hands should not move too far apart from one another, and consequently 



that one should not play too high with the right hand. This rule is reason- 
able and good, if only it could always be observed. It is a much better 
effect if the accompanying part on the keyboard (Fliigel) is taken below 
the solo part, than if it is taken in unison with the top part, or actually 
above it.' 

In this respect, too, exceptions can always be made for special effect. 
There is some evidence, moreover, that the preference for keeping the 
accompaniment below the solo was not as universal as Quantz implies. 
Walter Kolneder (Auffuhrungspraxis bet Vivaldi, Leipzig [1955], pp. 
85-9) gives examples of accompaniments, actually written out by 
Vivaldi, which move well above the solo part, and he suggests, I think 
correctly, that the Italian taste was more adventurous in this matter 
than the German. 

Ex. 183. Vivaldi, Concerto PV 31 1, Grave, solo violin accompanied 
by organ: 

fl Grave 

Bar II 

F T PT f if f'f 


.''V' JJ J. 

8 g. 5 6 

Among German composers, G. Ph. Telemann was somewhat excep- 
tional in taking his accompaniments quite freely above the solo part 
(cf. his examples in his Generalbass-Vbungen, Hamburg, 1733-35). He 
was also exceptional in having no objection to making a full close with 
the tonic 5 in the upper part of the accompaniment a small point on 
which many German authorities (including C. P. E. Bach, Essay., 
II, Berlin, 1762, II, i, 36) were sensitive. 

(d) When the solo part to be accompanied was itself at a low pitch, 
the accompaniment could freely move above it, but was expected to be 
in reasonable relationship to the solo. 

(357) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 22: 

'When accompanying an instrument of low pitch, such as a bassoon, a 
violoncello etc., or a low voice, Tenor or Bass, in a solo or aria, one must 
take care, in the matter of height, not to get too far from the solo part, 
but must pay attention to the register within which it moves. Here one 
seldom takes chords above the one-stroke octave [c 1 to c 11 ]. If it is necessary 
to take chords in a rather low register, the harmony should be thinned out, 



since a low pitch does not permit the harmony to be thick without losing 
its clearness. 9 

(e) When the continue part is given a clef which shows that it is no 
longer literally a bass line, but is reproducing some higher part in the 
temporary role of bass, all restrictions as to the height permitted to the 
accompaniment are waived. The clef will normally be one of the C 
clefs; occasionally a G clef. Such a bass passage is known as bassetto 
(Ger. Basset or Bassetgeri). Unfortunately, the appearance of a C clef 
is not quite a reliable indication of a bassetto passage, since it may be 
put merely for the convenience of avoiding ledger-lines when the 
normal bass-part goes somewhat high. But the effect is, with one 
exception, much the same: chords are necessarily taken higher up than 
usual. The exception is that a true bassetto passage is often the entry 
of an upper part in fugal music. Such an entry is not necessarily 
harmonised at aE: it may be taken senza cembalo, or tasto solo, or 
alYunisono (see 2 above). 

There was some variety of opinion as to whether a high bass part 
should have chords as full as usual, but higher up, or whether the 
chords themselves should be somewhat compressed to avoid an undue 
upward extension of the compass. It is obvious that either treatment, 
or a mixture of the two, may be most suitable to the passage in question, 
and that the accompanist should rely on his own discretion. If no 
figures are shown, or only a single line of figures, the probable intention 
is one or two parts at most (doubling one or two fugal entries without 
harmonising them). 

(358) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. V, 
Sect. 18, p. 516: 

'Bassetti must never be doubled with the octave below . . . 

'The chords over the Bassetto may, however, be played with both 
hands (yet with a sensible regard for the number of the voices and instru- 
ments taking part) in as many parts as there is room for in the space 
remaining at the top of the keyboard.' 

(359) Joachim Quantz, Essay, Berlin, 1752, XVII, vi, 27: 

* When the bass deserts its proper register and has something to play in 
the register of the tenor . . . the accompaniment should be kept in few 
parts and quite close to the left hand/ 

(f ) Both in general and in relation to the other parts, the accompanist 
can allow himself considerable discretion as to the compass of his part, 
provided that he respects the principle behind these rules, which is that 
the solo shall not be obscured but set off to the best advantage. The 
danger of going unusually high is that this is very likely to confuse the 
solo melody or melodies; if it does not do that, and if it makes a 



convincing effect of scoring, there Is no objection. Like any other special 
effect, however, It should not become habitual. 



The remaining, more structural aspects of the texture of the 
accompaniment will be considered in the following chapter. 



The Structure of the Part 


This chapter continues the discussion of the texture, but in its more 
structural aspects: the smoothness or otherwise of the part- writing; 
varieties of broken chords and other figuration; the degree of melodic 
independence and contrapuntal workmanship. 


(a) Just as four parts are the standard number on which other 
textures are variants, so the smooth style is the foundation for the other 
varieties of structure. 

The rules for the smooth style of accompaniment are two: any notes 
which can be held over into the next chord are held over, preferably 
without repercussion; and the position of each chord is taken as near 
to the position of the previous chord, and with as few and narrow leaps, 
as possible. 

(360) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. Ill, p. 19: 

'What decides the arrangement of the parts is the position of the accom- 
panying [Le. right] hand. The rule is, that when you have once placed the 
hand on the keyboard, to play the first chord of the melody you are going 
to accompany, you must take all the subsequent chords in the nearest 
possible place [with the result that] the parts change their arrangement 
with each chord you play. 

[Ch. V, Sect. 21, p. 86:] 'When one moves from one chord to another, 
one should see if any notes of the chord which one is leaving can be used 
in the chord which one is approaching; when that can be done, one should 
not change these notes. 

[Ch. V, Sect. 2, p. 66:] 'The right hand ought always to take its chords 
in the nearest position in which they are to be found, and never go far 
away in search of them . . . thus the accompanist always makes the chords 
progress by the smallest intervals possible.' 

Both for technical and for musical reasons, these rules are valuable. 

For technical reasons: because the right hand must be trained to 
take groups of figures as physical patterns of the fingers, this being the 
only way in which it can possibly get there in time, when one is sight- 
reading; and the less the hand jumps about the keyboard, the more 
readily it will find the positions into which these patterns fit. 



For musical reasons: because one of the primary functions of a 
continuo part Is to cement the harmony and bind the melodic parts 
together, by providing a steady flow of sound in the middle registers; 
and if it is jerkily realised, it will lose much of this binding influence. 
A part which offers the technical convenience of lying comfortably 
under the hands is also likely to be a part well designed for the musical 
purpose of supporting the existing melodies with a sustained sonority. 

(361) C P. E. Bach, Essay, II, Berlin, 1762, Introd., 25: 

'I shall treat of good construction and the smooth progression of the 
harmonies . . . avoiding unnecessary leaps and clumsy part-writing. 

[XXIX, 3 :] * A noble simplicity in the calm accompaniment . . . 

[14:] 'When playing chords over a bass not marked staccato, there is no 
need to make a fresh repercussion on each note. Notes which are already 
part of the previous chord and can be carried into the next are sustained; 
for this method, together with flowing progressions well layed out, gives 
the part a singing quality. 

[XXXII, 8:] 'Sustained notes . . . connect chords, help a cantabile 
style, are easier and less risky than repeated notes, which in four parts and 
at a rapid speed are nearly impracticable and in any case ineffectual/ 

(b) For the smooth style of accompaniment, it will be appreciated 
that the touch itself should be smooth. 

(362) Lorenzo Penna, Li Primi Albori, Bologna, 1672, Ch. XX, 
rule 8 : 

'It is always excellent to play legato, so as not to distract from the vocal 

(363) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 5: 

'Some people fall back on a very staccato touch when they want to 
achieve a piano [on a single-manual harpsichord without pedals, which 
cannot achieve it by change of stops in the middle of tlie piece], but this 
has a most harmful effect on the performance . . . it is better to decrease 
the volume by striking right-hand chords less frequently over passing 
notes [in the bass]/ 

(c) Even in the smoothest style, of course, the smoothness of the 
touch is always subject to the requirements of good phrasing and good 
articulation. The relatively short duration of harpsichord or lute tone 
(as opposed to the organ, or even the modem piano) has also to be 
born in mind, together with the need for rhythmic sharpness at certain 
points in the music, or as a help to good ensemble when there are 
many players. 

For the bass ties mentioned by Caccini in (364) below, see Ch. 
XXVII, 3 (c) above. 



(364) Giulio Caccini, Euridice, 1600, preface: 

1 have used the ties in the bass part because, after the consonance [or 
other interval figured] only the interval figured is to be struck again [and 
not the bass note itself] . . . for the rest, I leave it to the discretion of the 
more intelligent class of performer [on the lute] to strike again, with their 
bass, such notes as they judge to need it, or as will the better accompany 
the solo singer/ 

(365) Heinrich Albert, Arias 9 Part II, Konigsberg, 1640, preface, 
rule 5 : 

*[Tied notes should be sustained on the organ; but on the] harpsichord, 
as likewise on the lute or the bandora, since the sound of a string, once 
touched, soon fades and becomes weak, the fingers should be lifted, and 
both the suspensions and the consonances often repeated and struck, [not 
all at once but] so that now the top part, now an inner part and now the 
bass moves and serves its function/ 

(366) C. P. E. Bach, Essay, Berlin, 1753, Introd., 9 9 f.n.: 

It is thought quite proper and desirable for the accompanist to repeat 
chords which are held on by the other performers with a view to keeping 
up a clear indication of the measure. 

[II, 1762, XXEK, 18:] 'In a concerto or other thickly scored music, 
when the bass and ripieno parts sustain a note [chord] during which the 
solo part continues with its own movement, perhaps varying it occasion- 
ally with syncopation, it is advisable for the accompanist to keep the beat 
going and give a lead to the other performers by striking a chord in the 
right hand at the bar-lines even though the harmony does not change. If 
only the bass sustains the note, then the accompanist may repeat the bass 
note only just as it fades away; but this must not be done across the beat, 
as the saying goes. In a duple-time bar, the repetitions may come at the 
beginning and in the middle, according to time-unit and speed. In a 
triple-time bar, only the first beat is to be played. [Or at a sudden forte on 
any beat.] 

'[XXXVI, 14: Repeated pairs of bass notes of which the first is dotted] 
sometimes, for a particular kind of expression, require a [repeated] chord 
on each, instead of the more customary continuance of the same chord 
[unrepeated] over the short notes [after the dot]/ 


(a) One very important way of breaking the smooth style of accom- 
paniment is by arpeggiating: taking the notes of the chord in a free 
rhythm and any desired order and pattern. This, on the lute or the 
harpsichord, has the effect of keeping the sound going, and can thus 
actually be more sonorous than the smooth style itself. 

When the rhythm is not free but measured, the effect can be used in 



many contexts where free arpeggiation would be distracting. This 
measured effect is what we usually call broken chords. 

In both methods, as many notes are held on as the harmony, the 
phrasing and the available fingers permit. This gives the harpsichord 
some of the build-up of tone which the piano gets from its sustaining 
(damper-raising or loud') pedal, and is a crucial element in its tech- 
nique. Within smaller limits of volume, the same applies to the lute; 
and the name of 'lute-play* was often given to the arpeggiated style, 
even on the harpsichord. 

(367) Lorenzo Penna, LiPrimi Albori, Bologna, 1672, Ch. XX, rule 

'Take care to arpeggiate the chords so as not to leave empty spaces on the 
instrument [here used, as quite often at this date, to mean "plucked 
keyboard instrument"]/ 

(368) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. IX, Sect. 14, 
p. 132: 

'On the organ one does not repeat the chords, and one uses scarcely any 
arpeggiation: on the contrary one ties the sounds a great deal' 

(369) Michel de Saint-Lambert, Traitf, Paris, 1707, Ch. IX, Sect. 8, 
p. 130: 

'[Chords may be arpeggiated one or more times, upwards, downwards, 
etc.] But this reiteration, which wants to be very well managed, cannot be 
taught by book; it must be observed in practice. 

'[Unmeasured] arpeggiation is only suitable in [unmeasured] preludes, 
where there is no strict measure; for in airs with [measured] movement 
(Airs de mouvement) it is necessary to strike the chords definitely in time 
with the bass, except that when all the notes of the bass are crochets, in 
triple measure, one separates the notes of each chord in such a way as 
always to keep back one to sound between two beats/ 

Ex. 184. Saint-Lambert, loc. cit., broken chords: 

* a ^T 1 
.), i j i 


6 6 


i / .1 1 j j^^ 


There is actually no reason to restrict this technique to any one time 
or measure, and it is at once one of the simplest and one of the most 
useful forms of broken chord. It does not need to be the top note 
which is held back; an inner note can be equally effectual. Again, more 
notes than one can be held back. In a fuller accompaniment, part of 
the chord, or even the whole of it, can be brought on the beat by the 
left hand, and the remainder of the chord, or a reduplication of some 



or all of it can be held back. Indeed, the varieties are only limited by 
the necessity of sounding the bass note itself on its proper beat. 
Other typical methods of breaking the chords are shown in the 

following examples. 

Ex. 185. Johann Mattheson, Organisten-Probe, Hamburg, 1719, 
broken chords: 

n & m 

,S f A 

I 7 6 5\> 

t* t_ r f r - 

* 8 J 

p f: RP 

7 5 

-f F -F 

In Ex. 185 (c) above the mixture of melodic notes with harmonic ones 
will be observed. On the analogy of figurate arpeggios, we may call this 
figurate broken chords. Ex. 186 (a) below shows a fuller handling with 
the harmony sounded in plain left-hand chords against broken right- 
hand chords; Ex. 186 (b) shows the bass itself broken (where the rule 
must be observed to start the arpeggio on the actual bass note); Ex. 
186 (c) shows broken chords in either hand. 

Ex. 186. J. D. Heinichen, General-Bass, Dresden, 1728, Part I, 
Ch. VI, Sect. 3lff. 9 p. 556, broken chords: 

Bass as written 



Whether the broken chords are actually written into the bass part, or 
whether they are introduced there by the player, there is no need to 
put into the right-hand (unbroken) chord an interval both figured and 
present in the left-hand broken chord. Some authorities, ignoring the 
near approach to consecutive octaves which results, encouraged this 
doubling; others disallowed it. The choice depends partly on taste, 
partly on the fullness of the accompaniment. 

(370) David Kellner, General-Bass, Hamburg, 1732, 1, xvi: 

'Sometimes the bass makes a sort of invasion into a middle part; but 
the accompanist should keep to his normal accompaniment without 
taking any notice.' 

(371) G. P. Telemann, Generalbass-Ubungen, Hamburg, 1733-35 
No. 21 : 

'In the case of broken notes, that is to say where the bass includes one 
or more intervals which really fall within the province of the right 
hand . . . these may, indeed, be taken in one of the middle parts with the 
same progression [i.e. in consecutive octaves with the bass, concealed 
only by slight retardation], but we can scarcely allow this in the upper 
part; in view of this, we have preferred to leave out the 7 which in any 
case is sounded in the bass, rather than introduce two 8ves, namely b 
and #0.' 

Ex. 187. G. P. Telemann, loc. cit. 9 1 figured (but present in left-hand 
broken chord) not doubled in right hand : 

(372) C. P. E. Bach, Essay, II, Berlin, 1762, XXXIV, 2: 

'Where dissonances are resolved by a broken bass, the accompaniment 
[as figured] must be altered to avoid the disgusting retarded octaves. The 
dissonances included in the figures may therefore be unhesitatingly 
omitted [as so often, C. P. E. Bach is stricter here than most].' 


(373) Michel de Saint-Lambert, TraM, Paris, 1707, Ch. IX, p. 131 : 

'When accompanying a long recitative, it is sometimes good to dwell 
for a long time on one chord, when the bass allows, and to let many notes 
be sung by the voice without [audible] harpsichord accompaniment, then 



strike again a second chord, and next stop again, and thus only make an 
accompaniment at long intervals, assuming that as I have said the bass 
only has long notes, which is normally the case in recitative. 

*At other times, after striking a full chord on which you dwell for a 
long time, you strike one note again here and there, but with such good 
management that it seems as if the harpsichord had done it by itself, 
without the consent of the accompanist. 

*At other times again, doubling the intervals, you strike all the notes 
again one after the other, producing from the harpsichord a crackling 
almost Hke musketry fire; but having made this agreeable display for 
three or four bars, you stop quite short on some big harmonious chord 
(that is to say, without a dissonance) as though to recover from the effort 
of making such a noise.' 

(374) Nicolo Pasquali, Thorough-Bass Made Easy, Edinburgh, 1757, 
p. 47: 

'[The art of accompanying recitative] consists in filling up the harmony 
as much as possible; and therefore the left hand strikes the chords in it as 
well as the right. 

'Care must be taken not to strike abruptly, but in the harpeggio way 
laying down the fingers in the chords harp-like, i.e. one after another, 
sometimes slow, other times quick, according as the words express 
either common, tender, or passionate matters : 

'For example; for common speech a quick harpeggio; for the tender a 
slow one; and, for any thing of passion, where anger, surprise, etc., is 
expressed, little or no harpeggio, but rather dry strokes, playing with 
both hands almost [but, N.B., not quite] at once. 

'The abrupt way is also used at a punctum or full stop, where the sense 
is at an end/ 

For a full illustration of Pasquali's recommendations, see App. II 

(375) C. P. E. Bach, Essay, II, Berlin, 1762, XXXVIII, 3: 

'Arpeggiation is to be avoided in rapid declamation, especially when 
there are frequent changes of harmony. ... [It should be kept for] slow 
recitative and sustained chords/ 

For other aspects of recitative accompaniment, see Ch. XXXII, 
2 below. 


(a) A further variety of accompaniment is that in which the top part 
is not merely melodious in the sense of being flowing, but is an actual 
melody. This can happen in either of two ways, or a combination of 



both. The accompaniment can double the existing melody; or it can 
introduce an independent melodic interest of its own. 

(b) The early baroque method of accompanying polyphonic music 
by a reduplication of the existing polyphonic parts (see Ch. XXVII, 1 
above) gives the accompaniment melodic interest, but not independent 
melodic interest. 

This method, which is really the continuation of a renaissance 
tradition into the early baroque period, was still favoured by a minority 
of accompanists for some generations longer. The German organists 
of the seventeenth century, who in some respects were unusually con- 
servative, often insisted on an old-fashioned tablature from which they 
could follow and largely reduplicate the movement of the polyphony, 
The English chamber music of the viols during the greater part of the 
seventeenth century was an almost complete survival of sixteenth 
century polyphony in its construction, though ahead of its own times 
in the warmth and often in the boldness of its harmony; its written-out 
organ accompaniments are usually somewhat thinned reduplications of 
the string parts. 

In more progressive circles the method went rapidly out of fashion. 
It was never applicable to the monody of Caccini, Peri, Cavalieri or 
Monteverdi, as their use of compound figures makes it possible for us 
to be certain. Elsewhere, it was soon rejected, at first as unnecessary, 
and then as undesirable, by the leading theorists. 

(376) Agostino Agazzari, Del sonar e sopra 'I basso, Siena, 1607, end: 

'In conclusion, there is no necessity for the accompanist to play the 
parts in their existing form, so long as accompaniment is his purpose, and 
not a performance [i.e. an organ transcription] of the music itself, which 
is quite another subject.' 

(377) Michael Praetorius, Syntagma, III, Wolfenbiittel, 1619, Ch. 

It is not necessary for the organist in his accompaniment to follow the 
vocal parts as sung, but only for him to play his own version of the 
harmonies on the continue.' 

(378) Andreas Werckmeister, Anmerckungen, Aschersleben, 1698, 
ed. of 1715, Sect 69: 

'One must avoid continuous movement in octaves with the singers and 

(379) Francesco Gasparini, UArmonico Pratico, Venice, 1708, Ch. X: 

'One must never accompany note for note as in the voice part or any 
other top part composed for the violin, etc., since it is enough that the 



consonance or dissonance composed, or required by the bass, shall appear 
in the body of the harmony according to the rules of accompaniment.' 

It must be clearly understood, therefore, that the accompaniment 
should not ordinarily reproduce the melody accompanied. 

Apart from the exceptions mentioned above, there are a few situa- 
tions which can best be met by such reduplication, though not usually 
for long at a time. In other situations, the effect is dull and ponderous, 
and should be avoided. But reduplicating the lower of two solo parts 
in the middle of the accompaniment, either at the unison or at the 
octave below, is not so conspicuous, and may often be useful and not 
easy to avoid. 

(380) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 24: 

'Objectionable as is the accompaniment in which the upper line con- 
tinuously reduplicates the melody of the main part, it is nevertheless 
sometimes necessary, and therefore permissible, at the start of a rapid 
piece, particularly if this is in two parts. By this means the performers 
join hands over the tempo, and the listeners will miss no part of the 
opening, which will be kept unanimous and in good order. And generally 
speaking, unreliable musicians whether accompanists [here meaning 
ripienists] or leaders are allowed this assistance even apart from the open- 
ing, if they can thereby be helped to regain a lost steadiness of tempo.' 


(a) If the reduplication of the existing upper melody is to be avoided, 
the question remains how much independent melody is to be given to 
the accompaniment in place of it. 

The answer to this question depends largely but not entirely on 
circumstances. The remaining factors are, of course, the individual 
taste and talents of the performers. But it is only by adapting himself 
ungrudgingly to the circumstances of the case that the performer can 
indulge his taste and his talents without spoiling both the parts which 
he is accompanying and his own accompaniment. 

(b) The principle, as usual, is to put the interests of the existing 
melodies first. 

(381) Lodovico Viadana, Concerti Ecdesiastici, Venice, 1602, pre- 
face, rule 2: 

'The organist is obliged to play the organ part in a straightforward 
manner, especially in the left hand; if he wants, however, to introduce 
some movement into the right hand, for example by embellishing the 
cadences, or by some suitable free ornamentation, he must perform this 
in such a way that the singer or singers shall not be covered or confused 
by too much movement.' 



The reader is again referred back to the third paragraph of (129) In 
Ch. IX, 3 above, where Agazzari in 1607 develops this theme at 
somewhat greater length. 

(c) The most obvious application of this principle of putting the 
interests of the existing melodies first is that the accompanist should 
Ind the chief outlet for his own impulses towards elaboration at pre- 
cisely those points in the music where there is least elaboration in the 
solo parts. 

In passages where the continue is playing alone, the accompanist 
not only may but must produce his own elaboration to fill the gap. 
Such passages may occur at introductions before the solo entry, or as 
interludes when no soloist is for the moment being heard. 

(382) Lorenzo Penna, LiPrfmi Albori, Bologna, 1672, Ch. 14: 

'In the Ritomelli, or in the pauses intended to rest the singer, the 
organist must perform something of his own invention, in imitation of the 
Arietta or other lively matter which has just been sung.' 

(383) Friedrich Erhard Niedt, Musicalische Handleitung, Hamburg, 
1700, Ch. VIII, rule 8: 

'[Quick basses ordinarily need a chord] only with the first note of a 
half-bar or crotchet : the remaining [bass notes] are called passing notes . . . 

'But if it should chance that [an opportunity for] a solo passage were 
written into the continue, then the organist must play in a more florid 

The same kind of opportunity also occurs, though not quite so 
freely, when the solo part comes to any long sustained note against 
which the accompaniment can and should move more rapidly and 

(384) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. VI, 
Sect. 28, p. 547: 

'[An ornate accompaniment,] for which the cantabile solo and the 
unaccompanied ritornello of arias without [ripieno] instruments afford 
the best opportunity, [which can be fully exploited] so long as no harm is 
done thereby to the composed parts.' 

We see from this that not merely isolated long notes, but entire pieces 
in a cantabile style (Le. with a preponderance of sustained notes, not 
necessarily all of them particularly long) admitted a suitably ornate 
accompaniment. Experience confirms this, though not necessarily in 
the form of independent melody: broken chords and figurate arpeggios, 
for which see 3 and 4 above, may often provide the necessary movement 
and interest in a less distracting form. 



(385) C P. E. Bach, Essay, II, Berlin, 1762, XXIX, 3: 

'Where the opportunity arises and the nature of the music allows, 
when the solo part is resting or is performing plain notes, the accompanist 
may relax the control on his held-back fire. But this needs great ability 
and an understanding of the true feeling of the music. 

[4:] 'It is sometimes a necessary and indeed a perfectly proper course for 
the accompanist to discuss the music with the soloist before performing it, 
and let him decide how much liberty is to be taken with the accompani- 
ment. Some prefer their accompanist to be very restrained; others the 

[XXXII, 11 :] "Transitional passages offer a challenging invitation to the 
inventiveness of an accompanist. But his inventiveness must be in sym- 
pathy with the feeling and substance of the music. If some reminiscence of 
a previous phrase can be introduced, even at the cost of slightly altering 
the bass part and modifying the transition, so much the better ... so 
long as the solo part is not subjected to interference.' 

A small but very important variety of the transitional passages 
referred to in the last paragraph above occurs at the joins between 
sections, where two workings of the transition are often required: one 
leading back to the original key of the section for the repeat, the other 
leading to the key in which the next section opens. For examples, see 
Ch. XXVIII, 3 (c) above. 

(d) When the bass itself is florid, it is most unlikely that the remainder 
of the accompaniment should be florid too, except where the continuo 
is on its own. 

(386) Johann Mattheson, Organisten-Probe, Hamburg, 1719, Mittel- 
Classe Prob-Stiick 9, Erlauterungen, Sect. 1 : 

'There is seldom opportunity for ornamental and florid playing when 
the bass itself is purposely written in an ornamental and florid style. If, 
however, the bass is without particular embellishment, and one is playing 
alone, either in the introduction to an Aria, in the middle, or wherever 
else there may be a rest [for the soloist], then and there it is that these 
adornments, these figurations, these inventions, these embellishments 
find their proper, indeed one might almost say their necessary place/ 

(e) Among the embellishments which may be introduced into any 
part of the accompaniment, including the bass, are the specific orna- 

(387) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. IX, Sect. 15, 
p. 132: 

'One can on the organ, as well as on the harpsichord, make from time 
to time certain trills or other ornaments, in the bass as well as in the 
other parts.' 



(388) J. D. Heinichen, General-Bass, Dresden, 1728, Part 1, Ch. VI, 
Sect. 2, p. 521: 

"The art of a properly embellished continue [includes] sometimes 
introducing an ornament in all the parts (particularly the outside part of 
the right hand, which is the most conspicuous)/ 

(f) The extent to which opportunities can be found for independent 
melody in the accompaniment without detriment to the existing parts 
is not entirely predetermined, since it depends to some extent on the 
performer's ability to invent a counter-melody so exquisitely adapted 
to the existing melody as to enhance it rather than draw attention away 
from it. 

The following gives us some account of J. S. Bach's abilities in this 

(389) Lorenz Mizler, Musikalische Bibliothek, Leipzig, I, Part 4, 1738, 
p. 48: 

'Whoever wishes to form a real conception of refinement in contmuo, 
and of what good accompaniment means, has only to put himself to the 
trouble of hearing our Capellmeister [J. S.] Bach here, who performs any 
continuo to a solo so that one imagines that it is a concerted piece, and as 
if the melody which he plays in the right hand had been composed 
beforehand. I can bring living witness to this, since I have heard it myself/ 


(a) The independent melodic interest of the accompaniment may 
occasionally extend to taking equal part in the existing structure by 
way of imitation. 

(390) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. IX: 

'When one is accompanying a single voice which is singing some air 
with [measured] movement (Air de mouvement) in which there are a 
number of melodic imitations, such as the Italian arias, one can imitate 
on the harpsichord the subject and the points of imitation of the song, 
making the parts enter one after the other. But this demands a supreme 
skill and it must be of the first excellence to be of any value/ 

(391) J. D. Heinichen, General-Bass, Dresden, 1728, Ch. IV, Sect. 
40, p. 578: 

Imitation does not depend on your own invention, but must be taken 
from the written composition itself. Thus what is here called imitation is 
when the accompanist tries to copy the beginning of a phrase or an 
invention of the composer in passages into which the composer himself 
has not introduced it. But since we have never to interfere with the singer 
or player with this phrase, where he has it himself, and since, on the other 




hand, the composer himself must be imagined to have put it into the 
passages where the beginning of an imitation is in keeping, and thus to 
have left the accompanist little room for imitation, from all this it is 
evident that this kind of embellishment on the harpsichord is the most 
wretched of all. 

'But since a few cases may be found here or there, in some piece where 
(especially in Cantatas and Arias without [ripieno or obbligato] instruments) 
there is room left for a skilful accompanist to imitate a phrase occurring 
in the continuo or the solo part, more frequently than the composer has 
done himself, the following example may serve by way of illustration; 
here a further feature is to be observed, in that the right hand tends to 
want to accompany the solo part in 3rds and 6ths, and in this way to form 
a sort of concerted duet with it. This kind of imitation is particularly 
successful in vocal pieces, and is the easier to achieve [impromptu] in that 
with [solo] chamber music and with operatic music the [solo] part ordin- 
arily written above [the bass] makes it possible to watch the singer closely, 
to keep out of his way, and to follow his lead/ 

Heinichen's example to the above, which is a beautiful piece of neat 
and telling workmanship, will be found complete in App. II below. 
There is no doubt that imitation can be a most attractive enrichment 
of the accompaniment in the not very frequent instances in which it 
can be brought in easily and naturally and without any forcing of the 
overall effect. It is obviously much harder to improvise than to prepare 
beforehand, and it is, on the whole, one of those recourses over which 
it should be possible to work a prepared part more closely than a more 
or less impromptu one. Nevertheless, a quick enough accompanist can 
sometimes seize his chance at sight. 

(392) Johann Friedrich Daube, General-Bass, Leipzig, 1756, Ch. 
XI, Sect. 12 9 f.n.: 

'The admirable [J. S.] Bach commanded [the elaborate style] of accom- 
paniment in the highest measure; when he was the accompanist, the solo 
was bound to shine. He gave it life, where it had none, by his abundantly 
skilled accompaniment. He knew how to imitate it so cunningly in 
either right or left hand, and again how to introduce so unexpected a 
counter-melody, that the hearer would have sworn that it had all been 
composed in that manner with the greatest care. At the same time, the 
regular accompaniment [i.e. chords realising the figuring] was very little 
cut down. All in all, his accompanying was always like a concerted parted 
part worked out most elaborately, and matching up to the solo part, so 
that at the right time the solo part was bound to shine. This privilege was 
also granted even to the bass, without interference to the solo part. 
Suffice it to say that anyone who did not hear him missed a great deal.' 

(b) Daube's reference to imitations introduced into the left hand as 
well as into the right is accounted for by the passage to which the 



quotation above is a footnote. This distinctly states that the itself 
may be altered if a successful point of imitation can be gained by so 

(393) Johann Friedrich Daube, General-Bass, Leipzig, 1756, Ch. 
XI, Sect. 12: 

4 [The elaborate style of accompaniment] arises: (1) When, in masterful 
alternation with [the simple style] one tries from time to time to introduce 
suspensions where the composer has not written them nor shown them 
in the figuring. (2) When the solo part has a rest: here one can sometimes 
bring in some melodic passages. (3) One may also move in 3rds or 6ths 
with the solo part. (4) When one tries to imitate the theme of the solo 
part, or further, at one's discretion, to let a countermelody be heard. 
(5) As it may sometimes happen, moreover, that even with a good solo 
part the bass is badly constructed, either because it could have brought in 
an imitation, but this has been left out from carelessness or inadvertence, 
or because where it might have moved [better] in quick notes or slow 
notes, the opposite has been done; in such cases the accompanist may be 
allowed the liberty of attempting a correction while his accompaniment 

Changing the bass so as to introduce an imitation presupposes either 
the absence of any other bass instrument with which the change might 
conflict, or harmonic compatibility between the written bass and the 
improvised point of imitation. In the second case we have an opposite 
method to the characteristic seventeenth century method of keeping the 
fundamental bass on the keyboard and letting the melodic bass instru- 
ment make any improvised divisions. 

(c) Imitations both strict and free between the solo part and the bass 
are, of course, often introduced in writing by the composer. Where the 
soloist makes either an ornament or some impromptu variation in one 
of his entries, the accompanist must be quick to imitate this ornament 
or variation in his next answer. He may also be the one to instigate such 
a variation, if his own entry precedes that of the soloist. 

(394) C. P. E. Bach, Essay, II, Berlin, 1762, XXXIII, 1: 

Imitation often occurs in passages of which the repeat is varied 
[impromptu]. The accompanist must take part in the variation so that his 
imitation shall remain recognisable and shall not fall short in beauty. The 
accompaniment must follow the pattern of the solo part as closely as 
possible. [2:] In [Ex. 188] the bass leads the imitation. 

[5 :] 'A keyboard player who has other bass players doubling him must 
refrain from variations if he is doubtful whether the others will follow 

6 [6: Of imitation added to the accompaniment by the performer, in 
addition to that given by the composer; sometimes] neither the bass nor 



the accompaniment can imitate the solo part exactly. In such cases it is 
sufficient to invent an imitation of which the [rhythmic] outline matches 
that of the solo part, while the bass is left with its notes as they are or 
slightly varied. 

[7:] 'Those with good knowledge of part-writing can furthermore 
bring into a middle part an ornamental imitation of the solo part instead 
of the ordinary accompaniment. . . . The speed must not be very fast, 
however, or the effect will be confused.' 

Ex. 188. C. P. E. Bach, loc. cit. 9 (a) imitation started by the bass; 
(b) the same imitation with a variation, also started by the bass, as 
it might be improvised in performance during the repeat: 


L-L-J 1 

(d) Apart from imitation of any kind, a certain amount of neat 
contrapuntal workmanship is often very effective in the right context. 
The more formal the music, the more likely such contrapuntal work- 
manship is to suit it well; but in all cases, it will be appreciated that 
this workmanship should not sound stiff or self-conscious. There can, 
indeed, be a very light and delicate counterpoint, often in no more 
than two parts (including the bass), which can give a more exhilirating 
effect in a crisp and rapid movement than any succession of chords no 
matter how thin. At the opposite extreme there stands the solid crafts- 
manship of a movement in the learned' style. There is, in short, 
counterpoint to suit all calls and all situations, and while by no means 
all movements will tolerate any kind of counterpoint, it is surprising 
how frequently the more informal kinds can add a certain cleanness to 
the technique and a certain definiteness to the idiom. The part will gain 
in tautness and the accompanist will sound as if he knew exactly where 
he was going. When he relaxes into a fanciful and casual passage, he 
will gain all the more effect from the contrast. Such variety is the life 
of good accompaniment. 



Instruments of Accompaniment 


(a) At the start of the baroque period, the choice of accompanying 
Instruments was wide and flexible. 

In church music, the organ had pre-eminence; many performances of 
music once thought to have been always unaccompanied evidently 
included it, but as a reduplication of the polyphony, not as an inde- 
pendent texture. The range of melodic Instruments which might be 
introduced was also extensive. These, too, doubled the polyphony; but 
as the monodic style came into sacred use, they required parts of their 
own, which were not at first always provided by the composer. 

In secular song, the lute was pre-eminent during the late renaissance, 
and remained so at the beginning of the baroque period: particularly 
Caccini's favourite, the large accompanying arch-lute or chitarrone, 
which took a conspicuous share in the early monody. It could be used 
alone; but many options were open to the performers, and here again 
a number of melodic accompanying instruments might be assembled, 
the parts for which had normally to be more or less improvised in 
performance in so far as they had not already been preconcerted in 

These conditions of performance fall somewhat outside the main 
problem of baroque accompaniment as here understood, and have 
already been discussed, as a problem in free ornamentation, in Ch. 
IX, 3 above. The reader is particularly referred to the long quotation 
given there at (129), showing Agazzari's mention of various accompany- 
ing instruments. 

(b) It is, however, worth noting here that such contemporary lute 
accompaniments to early monodic songs as have survived in fully 
written-out form (e.g. Caccini's famous 'Amarilli' in Robert Dowland's 
Musicall Banquet, London, 1610) are conspicuously simple: much more 
so, for example, than John Dowland's characteristically contrapuntal 
settings of his own songs (Robert Dowland was the son of John). This 
is in full accordance with the insistence of the Italian monodists and 
their imitators that nothing in the music should be allowed to detract 
from the direct dramatic impact of the words; but in view of the com- 
plexity of melodic accompaniment described by Agazzari and Praetorius 



(see Ch. IX, 3, above) it is impossible to believe that lutenists and 
keyboard players, even when accompanying monody, were always 
reticent, and we may allow something for the usual diversity of 
individual practice. 

(c) In the early baroque period, the harpsichord, though carried to 
great lengths of virtuosity as a solo instrument, was not yet exploited 
with the same virtuosity as an instrument of accompaniment. It will be 
remembered that in Diego Ortiz' Tratado of 1553 it is the viol which 
makes the elaborate divisions, the harpsichord which sustains the solid 
part- writing of the original motet or chanson; and the roles are not yet 
reversed in early baroque music. A few momentary points of imitation, 
an occasional ornamental flourish and a considerable recourse to broken 
chords are the only relief to straightforward part-writing which should 
ordinarily diversify an early baroque accompaniment on the harpsi- 

(d) The organ, apart from its massive share in church music, was 
during the renaissance, and remained into the baroque period, an 
important instrument in chamber music: not in its larger representa- 
tives, but in the smaller positives and portatives, and in the little regal 
with its surprisingly telling and incisive tone, produced by beating 
reeds with very short tubes. 

The evidence suggests that an organ part for accompaniment would 
be still more regular in its part-writing than a harpsichord part, and 
still more subject to the accepted ideal of reduplicating the polyphony, 
when present, and otherwise of replacing it by a similar texture. 


(a) At the very beginning of the baroque period, and at the same 
time, therefore, as the colourful ensembles in the traditions of the 
renaissance mentioned by Agazzari, the emergence of music supported 
by a true 'general bass' was laying down the pattern for the most 
characteristic of all baroque combinations for accompaniment: an 
instrument (or instruments) of harmony in conjunction with a melodic 
bass instrument (or instruments) to reinforce the bass line itself. 

(395) Michael Praetorius, Syntagma, III, Wolfenbiittel, 1619, Ch. 
VI, Sect, on Organ, App.: 

'It is further specially to be noticed that when 2 or 3 voices sing accom- 
panied by the General Bass which the Organist or Lutenist has in front of 
him and from which he plays, it is very good, and indeed almost essential, 
to have this same General Bass played in addition by some bass instru- 
ment, such as a bassoon, a dolcian [a form of bassoon], or a trombone, or 



best of all, on a violone [probably meaning here a bass, rather double- 
bass viol].' 

The entire structural tendency of baroque music encouraged this 
disposition of forces. However important the middle parts, it is rare, 
in a characteristic piece of baroque music, for them to equal the outer 
parts in importance. It is this fact which makes the system of figured 
bass so eminently suitable for its purpose. The composer establishes the 
upper melody or melodies; he supports them with an equally strong 
melodic bass; but what lies between these outer extremes is of a 
different texture, and is not meant to be so sharply defined. These 
requirements can only be met with complete satisfactoriness when the 
bass line is itself outlined by an instrument or instruments of melody. 

Voice and lute, recorder or violin with harpsichord can be perfectly 
satisfactory. Trio sonatas without a melodic bass are apt to be much 
less so. The bass should, moreover, be played with about as much 
volume and significance as the solo parts, actually standing out above 
them for fugal entries and other prominent matter. It will not serve its 
function if it is played in a discreet and retiring manner. 

J. S. Bach's sonatas for violin and obbligato harpsichord are, in effect, 
trios with one melodic line on the violin and one in each hand on the 
harpsichord, with frequent interchange of themes. A partially auto- 
graph MS. calls them 'Six Sonatas for Harpsichord Obbligato and 
Violin Solo, with Bass accompanied on the Viola da Garnba at will'; 
but here the effect is somewhat to unbalance the fine distribution of the 
parts, and is hardly desirable. 

Titles such as Thorough-Bass for the Harpsichord or Violoncello', 
etc.., are to be taken in the sense of 'and/or' : cases occur (e.g. Vivaldi's 
Op. I) with 'or' in the title but 'and' in the parts. Two bass parts are 
usual, both figured, but only because they are printed from the same 
plates to save expense: one is for the melodic bass. No doubt the 4 or' 
was for commercial reasons; it encouraged amateur buyers who might 
not muster the complete complement of instruments. In the same way 
they were often given the choice of violins, flutes, recorders or oboes 
as solo instruments. 

(396-7) [Peter Prelleur] Modern Mustek-Master, London, 1730, ed. 
of 1731, in *A Dictionary': 

'Organo, signifies properly an Organ, but when it is written, over any 
Piece of Musick, then it signifies the Thorough Bass.* 

(398) Sebastien de Brossard, Dictionaire de musique, Paris, 1703, 
etc., s.v. 'Basso-Continuo' : 

'We also often play it [the continuo] simply, and without figures 
[harmonies] on the Bass Viol [gamba] or the Bass Violin [violoncello], 



together with the Bassoon, the Serpent, etc. [as alternatives], whence the 
Italians also call it Basso Viola, Violone, Fagotto, etc." 

(399) Fr. Couperin, Legons de Tenebres, Paris, 1714, Avertissement: 

'Although their melody is notated in the treble clef, all other kinds of 
voice can sing them, seeing that most present-day accompanists know 
how to transpose. . . . If one can join a Bass Viol [gamba] or Bass Violin 
[violoncello] to the accompaniment of the Organ or the Harpsichord, 
that will be good.' 

(400) C P. E. Bach, Essay, II, Berlin, 1762, Introduction, 8: 

'Some people have themselves accompanied in a solo on the viola, or 
even on the violin, with no keyboard instrument. If this is done from 
necessity, for lack of a good keyboard player, they must be excused [but 
not otherwise]. 

[9:] 'The most complete accompaniment to a Solo, and the one to 
which no possible exception can be taken, is a keyboard instrument in 
combination with the Violoncello.' 

The effect of a melodic bass instrument with no instrument of 
harmony is more successful in trio sonatas, where there are three real 
parts, than in solos, where there are two. It sometimes sounds sur- 
prisingly complete and beautiful. 

As a substitute for the full continuo with both melodic and harmonic 
continue instruments present, either alone is quite justified on baroque 
precedents, but neither is nearly as satisfactory as the two together. 


For orchestral music, the cello was and is the obvious choice, except 
that as the concertante bass in a concerto grosso the gamba makes an 
attractive variant. 

For chamber music, Italy was probably the first country to develop 
a preference for the cello, and France the last not until more than 
half-way through the eighteenth century. The gamba led in PurcelTs 
England, and was still quite active in J. S. Bach's Germany. The 
following is a somewhat late reference. 

(401) Dr. Charles Burney, Present State of Music in Germany, 
London, 1773, 1, p. 139: 

'[The Elector of Bavaria] pkyed [the bass of] one of SchwindTs trios 
[trio sonatas] on his Viol da gamba, charmingly: except Mr Abel, I never 
heard so fine a player on that instrument.' 

(402) [J. S. Sainsbury] A Dictionary of Musicians, London, 1824, 
2nd ed. 1827 [KB. the date], s.v. 'Abel': 

'The viol da gamba, now hardly used.' 



(b) The gamba, being more lightly constructed and strung the 
cello, has a less massive but more transparent tone, with Its own 
peculiar warmth and beauty of colouring. It is nearer to violin tone 
than the cello; and the blend of two violins and gamba in a baroque 
trio sonata, with the harpsichord to give a crystalline edge and sparkle, 
is one of the best sonorities in music. 

The cello also makes a magnificent effect here, but a more bottom- 
heavy one. There is not so close a blend of tone. Care should be taken 
to avoid its more ponderous colourings. 

(c) A second main difference between gamba and violoncello is their 
technique of bowing, which was originally the same (underhand) but 
was changed by Italian violoncellists of the early eighteenth century 
(and subsequently elsewhere) into an approximation to the (overhand) 
bowing of the violin. The resultant technique is more forceful, and better 
suited to the violoncello; the gamba technique gives, however, a 
pointed articulation characteristic of this instrument, and in its different 
style equally effective. This difference should be retained, and not 
obliterated by using violoncello bowing on the gamba. Here, too, there 
is a noticeable distinction between the effect of giving the continue line 
to a gamba and giving it to a violoncello. Like so many such distinc- 
tions, it is valuable for its own sake, and adds to the subtle possibilities 
of varying the accompaniment. 


In larger chamber groupings, the gamba as a continuo bass is at its 
best when supported at the octave below by a double-bass instrument. 
Its own double-bass, the violone, which has the same relative lightness 
of construction and the same six relatively thin strings, makes with it a 
combined sonority of remarkable silkiness and tranquil depth. A 
normal modern double-bass (which is a double-bass violin modified by 
some refining influence from the double-bass viol) is also extremely 
satisfactory if it is played purely; the smaller 'chamber* basses being 
the most suitable for this purpose. 

In orchestral music, gambas on the bass line can be multiplied in the 
same way as cellos, and given the same support from the double-basses. 
(They can also be used as tenor instruments.) 


(a) A harpsichordist supported by a melodic bass instrument is 
relieved of the responsibility for making special efforts to bring out the 
bass line in sufficient strength. 



(b) The less support from other instruments, the more need for 
correctness in the harpsichord part. 

(403) Michel de Saint-Lambert, Traite, Paris, 1707, Ch. VIII: 

1 [Forbidden consecutives do not matter when disguised by the presence 
of many voices or instruments.] But in accompanying a single voice, 
particular when unsupported [by a melodic bass instrument], it is impos- 
sible to be too strictly correct; for in that case everything shows clearly, 
and the Critics will tolerate no fault.' 

(c) The bass-line can be simplified on the harpsichord by leaving 
florid divisions to the melodic bass instrument: see (311) in Ch. XXVIII, 
3 (b) above. Indeed, in PurcelFs trio sonatas of 1685 and 1697 we actu- 
ally find the continuo bass part separately printed; it is a slightly 
plainer version of the string bass part, lacking some of the detailed 
figuration and exact fugal imitation. The discrepancies sound better 
when the harmonic instrument is a chamber organ, less convincing 
when it is a harpsichord. 

When, as is normal, the two bass parts are printed in duplicate from 
the same plates, the choice in this matter rests with the player. In spite 
of the historical justification for sometimes doing otherwise, it is 
ordinarily by far the most satisfactory course for the harpsichord to 
double all the notes of the melodic bass instrument with the sole 
exception of rapid repeated notes for which see Ch. XXVIII, 3 (b) 


(a) The choice of instrument greatly affects the nature of the part. 

(404) C. P. E. Bach, Essay, II, Berlin, 1762, Introd., 1 : 

'The organ, the harpsichord, the pianoforte, and the clavichord are the 
keyboard instruments most commonly used for accompaniment. 

[3 :] 'The organ is indispensable in church music on account of the 
fugues, loud choruses, and more generally, on account of the binding 
effect. It adds to die grandness and it preserves order. 

[4 :] 'But whenever Recitatives and Arias are used in church, especially 
those whose inner parts [on ripieno instruments] accompany simply so as 
to leave the voice free to make ornamental variations, there must be a 

unfortunately we all too often hear how bare the performance of such 
music sounds without the accompaniment of the harpsichord. 

[5 :] 'This instrument is also essential in the theatre or in a room, on 
account of such Arias and Recitatives. 

[6:] 'The pianoforte and the clavichord make the best accompaniment 
in a performance associated with the greatest refinements of [galant] taste. 



However, some singers would rather be accompanied on the clavichord 
or the harpsichord than on the first-mentioned, 

[7:] *No piece, therefore, can be adequately performed without the 
accompaniment of a keyboard instrument. Even with music on the 
grandest scale, in opera indeed, even in the open air, when it would be 
confidently supposed that the harpsichord could not be heard at all, It Is 
missed if it Is not present. A listener who is above the performers can 
clearly distinguish every note. I speak from experience, and anyone can 
try the same experiment/ 

(b) The organ Is an instrument which can take such very different 
forms that it is hardly possible to generalise about Its uses for accom- 
paniment. However, two main aspects may be mentioned. 

As an accompaniment to large-scale church music we are, of course, 
familiar with the organ, and the only comment required here is that 
baroque organs are markedly different from modern organs (other than 
deliberate replicas) in their scaling and registers. Wind pressure is low, 
but voicing is very positive. The attack is sharp, the tone unforced but 
colourful, and employed more for strong contrasts than for smooth 
gradations. Pungent mixtures are another characteristic feature. Some 
excellent baroque church organs survive, mainly on the Continent; 
among modern replicas, the baroque department of the Royal Festival 
Hall organ, London, designer Ralph Downes, is of great interest. 

Small and often portable organs were used for chamber music. Some 
had a single rank, usually wide open wooden diapasons; others had 
several stops. Like the harpsichord, the chamber organ blends with 
other instruments, but without the sharp edge and brilliance: its effect 
with strings is well described by Thomas Mace in 1676 see (34) in 
Ch. II, 5 above 'the Organ Evenly, and Softly, and Sweetly Ac- 
chording to All'. The same passage makes it clear that in late seventeenth 
century England the chamber organ was at least as common an accom- 
panying instrument as the harpsichord: Mace performed livelier music, 
e.g. violin as opposed to viol music, 'not to the Organ (as many (now- 
adays) Improperly, and Unadvisedly perform such like Consorts with) 
but to the Harpsicon'. The evidence of part-books suggests that it may 
have been more common, unless the word 'organ' is there used loosely 
for 'continue instrument' as Prelleur stated that 'organo* was, in 1731 
see (397) in 2 above. 

The titles of Corelli's 'church sonatas', Op. I and III specify Basso 
per FOrgano and of his 'chamber sonatas' Op. II and IV Violone o 
Cembalo ('big viol and/or harpsichord'): a useful hint, but the choice 
remained at the performer's discretion. 

(c) The harpsichord, large or small, was the standard baroque instru- 
ment of accompaniment. Even the smaller forms provide a satisfactory 



continue if there is a melodic instrument on the bass line as well. With 
the larger harpsichords, it should be remembered that hand-stops, not 
pedals, were the normal means of changing registers on baroque 
instruments, and that this greatly limited the player's ability to change 
the registration in the middle of a passage. The pedals of a modern 
harpsichord are much more convenient, and in view of C. P. E. Bach's 
insistence at (326) in Ch. XXIX, 4 above on adapting the volume to that 
of the solo parts, we should not hesitate to make good use of them for 
this purpose (see Ch. LXIV, h below for his praise of one of the rare 
pedal harpsichords of his period, and Mace's praise of another in 1676). 
On the other hand, there is undoubtedly some temptation to use the 
pedals too much, and this should be borne in mind. Changes of regis- 
tration imposed on the music where it does not really lend itself to them 
can become very restless, and are against the basic nature of the 

Balance can be a problem. In chamber music, the harpsichord part, 
like the bass-line, is useless if it sounds too quiet to give proper support. 
It is not a mere discreet background; it is there to be heard. Broad- 
casting and recording balancers should bear this in mind. It is, however, 
also possible to hear too much harpsichord: i.e. if it actually distracts 
attention from the melodic parts. 

A harpsichord concerto needs the solo to be heard easily, but a 
continue harpsichord in the orchestra does not need to be very 
consciously heard, provided it can make its presence felt by the clarity 
and sharpness it imparts. Even so, it may have to be electrically 
amplified for a large hall and orchestra, though this always sounds a 
little unnatural. A better alternative when practicable, and one well 
justified historically, is to double the continue with two or more 

(d) The mid-eighteenth century piano included by C. P. E. Bach in 
his list of accompanying instruments at (404) above had a tone, though 
not an attack, nearer to the harpsichord than to the modern piano, 
which is unsuited to baroque accompaniment, not being a sufficiently 
blending instrument. Throwing most of its acoustic energy into its 
lowest few partials, it has a solid neutrality of tone which tends to 
obscure the colouring of the instruments it is accompanying, unless the 
dampers are raised by using the loud pedal', when the upper partials 
come surging in to brighten and colour the tone, but at the expense of 
the clarity so necessary in baroque music. 

It is, however, bad advice to use a dry touch and a staccato articula- 
tion under the illusion that this imitates the harpsichord which should 
sound neither dry nor as a rule staccato. It is much better to play with 
natural warmth of feeling a genuinely though not heavily pianistic 



accompaniment, thus making the best instead of the worst of an 
unavoidable handicap. 

(e) C. P. E. Bach's own clavichord was a Silberaiann of exceptional 
quality and power (J. H. Reichardt, cited by W. J. Mitchell, trans, of 
C. P. E. Bach's Essay * New York, 1949, p. 36n); even so it is surprising 
and interesting, to find him at (404) above taking the clavichord for 
granted as an accompanying instrument. Its dynamic range is wide, 
because of its extreme pianissimo; but it has a very small absolute 
volume for its fortissimo. 

(f ) The lute, not being a keyboard instrument, was not Included in 
C. P. E. Bach's list, but is mentioned elsewhere as an accompanying 
instrument at a considerably later date. 

(405) John Hoyle, Dictionary of Music, ed. of 1791, s.v. 'Arcileuto': 

'An Arch Lute, or very long and large Lute, with Bass Strings, like a 
Theorbo Lute. . . and is much used by the Italians for playing a Thorough 

[s.v. 'Bass (Thorough)':] *Is the harmony made by the Bass Viols, or 
Theorboes, continuing to play, both while the Voices sing, and the other 
instruments perform their parts . . .' 

Hoyle's dictionary is a mere popular compilation from Grassineau 
and Geminiani, and of little evidential value in itself. This point is, 
however, made again by Busby, a rather better authority. 

(406) T. Busby, Complete Dictionary, ed. of [1801], s.v. 'Arch-Lute*: 

4 A Theorbo, or Large Lute. . . . It is still used in Italy/ 

A lute is louder than a clavichord, less loud than a harpsichord. It 
is capable of the greatest refinement of expression and tone-colouring 
and is an especially satisfying accompaniment for a solo voice, with or 
without a gamba to double the melodic bass-line. Its carrying-power 
varies remarkably with the skill of the player in producing that easy, 
free vibration which is so much more telling than violence on any 
musical instrument (the voice included). But in a fairly large hall, 
electrical amplification is the only recourse. In early baroque orchestral 
music, on the other hand, twelve or more lutes might be assembled 
together: that should dispense with any need for electricity. 



The Good Accompanist 


The best of accompanists need not be too proud to make as much 
preparation as the situation allows and as he feels that it calls for. 

(407) Lodovico Viadana, Concerti Ecdesiastici, Venice, 1602, pre- 
face, rule 3 : 

*It will further be desirable for the Organist to take a preliminary look 
at the Concerto which is to be sung, since he will always make his accom- 
paniments better for understanding the music.' 

(408) Lorenzo Penna, LiPrimi Albori, Bologna, 1672, Book III, Ch. 
XX, rule 1: 

"Make sure, so soon as you are given the part from which you are to 
play, that you understand its character, so that you may accompany 

(409) Le Cerf de La Vieville, 'Comparaison', in Histoire de la Musique, 
Paris, 1725, 1, p. 297 (a prejudiced account but none the less a warning) : 

"All that is generally heard in the Italian music is a thorough-bass 
^ccompaniinent unceasingly varied (doublee), this variation being often a 
kind of breaking of chords, and an arpeggiation, which throws dust in the 
eyes of those who know no better, and which, reduced to its elements, 
comes back to the same as ours. These thorough-basses are good only for 
showing off the quickness of hand of those who accompany on the 
clavichord or the [bass] viol [gamba] or again, to go one better (rencherir) 
on these basses, already varied in themselves, they vary them further, and 
it goes to him who will vary them the most; in such sort, that you cannot 
hear the melody any more, which would seem too naked after so much 
brilliance, and vanishes buried (demeure enseveli) beneath a chaos of 
embroidered (tricottez) and rattling (petillans) sounds, which pass so 
lightly that they cannot make any harmony against the melody. There 
then have to be two instruments, one to play the plain version of the bass, 
and the other the variation. These thorough-bass accompaniments pass 
rather for solos for the viol, than for an accompaniment, which ought to 
be subordinate to the melody, and not overpower it. The voice ought to 
stand out and attract the main attention; just the opposite happens here: 
you only hear the thorough-bass accompaniment, which rattles so loudly, 



that the voice is smothered. An awkwardness in with broken 

chords and variations [improvised] on the spot (sur le champ) ; this is that 
it is difficult for a harpsichord, a [bass] viol and a theorbo [arch-lute] to 
contrive to come together correctly, in the same kind of variation, not to 
mention other string or wind instruments; the one takes one turning, the 
other another, and the result is a remarkable cacophony such that the 
composer can no longer recognise his own work, which comes forth 
totally disfigured. You are supposed, in the midst of all this, to take 
pleasure in admiring the quickness of hand of the performers. There you 
have, at any rate, the modern taste in performance of the Italian music* so 
greatly vaunted.' 

(410) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 2: 

'The accompanist is usually given much less time [than the soloist]; 
he is only allowed a brief examination of the music [of which he should 
make the best use he can]. 

[XXXII, 3:] 'The commonest phrase to describe a good accompanist is; 
"he accompanies with judgment". Such praise covers a multitude of 
qualities. It amounts to saying that the accompanist can discriminate, and 
therefore realise his part to suit the character of a piece, the number of 
parts, the remaining performers and especially the soloist, the instruments 
and voices concerned, the auditorium and the audience. ... He puts 
himself into sympathy with the intentions both of performer and of 
composer, and he tries to help and back up these intentions. He uses every 
refinement of execution and realisation provided it is in keeping with 
the emotional requirements of the music. But in using such refinements he 
takes the greatest care that they shall not interfere with anyone. With 
this aim in view he does not make extravagant use of his accomplish- 
ments, but spends them economically and only where they can be of 
good effect. He does not always have to know best, and he never forgets 
that he is an accompanist, and not a soloist. ... In short, an accompanist 
of judgment needs a soul full of fine musicianship, including great 
understanding and great good will/ 


(411) C. P. E. Bach, Essay, II, Berlin, 1762, XXXVIII, 2: 

'Some recitatives, in which the bass, and possibly other instruments as 
well, perform a well-defined melodic line or continuous progression 
which does not share in the singer's pauses, are to be taken in strict time so 
as to keep them in order. The rest are declaimed according to their con- 
tent, now slow, now fast, without regard to the measure, even though 
they are written with bars. In both cases, above all the latter, an accom- 
panist must be attentive. He must listen constantly to the soloist, and when 
there is dramatic action, watch him too, so as to be always punctual with 
his accompaniment, for he must never desert his singer. 



[3:] 'When the declamation is rapid, the chords must be ready in the 
instant, particularly at pauses in the solo part where the chord precedes a 
subsequent entry. When the chord ends, the next one must be played 
promptly. In this way the singer will not be held back in his expression or 
in the rapid delivery needed for it; for he will always be warned in good 
time of the progression and nature of the harmony. If it is necessary 
to choose between evils, rather hurry than hold back; but indeed better is 
always better. Arpeggiation is to be avoided in rapid declamation, 
especially when there are frequent changes of harmony. For one thing, 
there is no time for it, and even if there were time, it might very easily 
throw accompanist, singer and audience into confusion. Moreover, 
arpeggiation is not needed here; its natural use is in quite different con- 
texts, namely slow recitative and sustained chords. In such contexts it 
helps to remind the singer that he is to remain in the given harmony, and 
prevents him from losing his pitch through the length of the chord, or 
from concluding that the chord has changed. 

[4:] 'The rapidity with which a chord is arpeggiated depends on the 
speed and character of the recitative. The slower and more expressive the 
recitative, the slower the arpeggiation. Recitatives in which the accom- 
panying [ripieno] instruments have sustained parts are well suited by 
arpeggiation. But so soon as the accompaniment changes from sustained 
notes to short, detached notes, the accompanist must play short, firm 
chords, without arpeggiation, and grasped entire in both hands. Even if 
tied white notes are written in the score, resort must still be had to short, 
detached playing. 

[5:] 'The organ holds only the bass of recitatives accompanied by 
instruments with sustained parts; it takes off its chords soon after playing 
them . . . arpeggiation is not used on the organ. Apart from arpeggiation, 
the other keyboard instruments bring in no ornaments or elaborations 
when accompanying recitatives. 

[7:] 'When a singer departs from the written notes, it is better to play 
repetitions of a full chord rather than single notes. The right harmony is 
the main factor in recitative; singers should not be expected to sing 
nothing but the written notes, especially in passages carelessly composed. 
It is sufficient if they keep their declamation within the right harmony. . . 
The reason may be a wish to keep in a comfortable register, or merely a 
failure of memory. 

[8:] 'When completing the arpeggiation of a preparatory chord it is as 
well to reach the top of die arpeggio with the note on which the singer is 
to begin [even if this means irregularities in the progressions; but] gener- 
ally this can be achieved quite easily without any such liberties by means 
of a rapid arpeggio.' 


Whether committed to paper or not, the realisation of figured bass 
is a branch of composition, and one with the very interesting condition 



attached to It of working to a given pattern and with partly given 
material. An accompanist who finds himself making small, half- 
instinctive changes in an editor's part as he goes along, who thins a 
chord here or changes a detail of figuration there, has already set foot 
on a fascinating path. He is co-operating with genius, in the person of 
the composer; in however small a way, he is contributing to a master- 

It is a principle of composition that material should be used econo- 
mically. If a counter-melody is needed, it is uneconomical and distract- 
ing to use entirely new thematic material; something not necessarily 
imitated from the existing melodies, but at least derived from them, is 
more concise and effectual. If a good and suitable figure of accompani- 
ment has been established, we shall not change it unnecessarily before 
it has had time to make its proper effect. One thing must grow out of 
another; the contrasts must contribute to the underlying unity; the 
part must sound consistent both with itself and with the written 
melodies. The less the material, the more inventively we are compelled 
to develop it; and it is wealth of development, not wealth of material, 
which enriches a work of art. In these and other ways, we need a 
composer's ear for what has gone before and Ms quick grasp of what 
can most naturally develop out of it. 

Francesco Gasparini (UArmonico pratico al cimbalo, Venice, 1708, 
Engl. tr. p. 42) calls figured bass accompaniment 'improvised composi- 
tion'; and it ought to sound so. Hence it ought also to sound complete 
in itself, and not to depend for its completeness on the existing parts 
which it is accompanying. This is a most useful principle, which experi- 
ence over and over again substantiates. Good 'voice-leading' can make 
even the simplest accompaniment sound a convincing and satisfactory 
entity in its own right. It can then give adequate musical support to the 
existing parts. If it depends upon their presence to make it sound 
complete, it cannot. 

(412) Roger North, Brit. Mus. MS. Add. 32531, c. 1710, /. 29: 

' Altho a man may attain the art to strike the accords true to a thro-base 
prescribed him according as it is figured, yet he may not pretend to be 
master of his part, without being a master of composition in general, for 
there is occasion of so much management in ye manner of play, some- 
times striking only by accords, sometimes arpeggiando, sometimes 
touching ye air, and perpetually observing the emphatick places, to fill, 
forbear, or adorne, with a just favour, that a thro-base master, and not an 
ayerist, is but an abecedarian, besides that which is called voluntary, which 
is a ready touching an air, and perfection of Musick (as is expected) upon 
every key, will not be done without a mastery of composition/ 






Expression In Early Music 


Expression is the reaction between what a performer brings to the 
music and what he finds there. If he brings too untrained a conception 
of the style, he is bound to inflict any number of ingenious wrong 
interpretations on to it, simply because he does not know the right ones. 
If he has some training, he can put his knowledge out of his mind and 
let his own musicianship produce the right expression exactly as it does 
in the later styles with which he has grown up from childhood. But 
indeed a generation will be growing up soon which can take many of 
the early styles as much for granted as we take the nineteenth-century 

Marks of expression, as opposed to expression, seem (like signs for 
ornaments) to have been characteristically few in renaissance music; 
nor did they (as the signs for ornaments did) proliferate during the 
baroque period. They became gradually commoner through the eigh- 
teenth century, plentiful in the nineteenth, almost exaggeratedly so in, 
for example, Mahler, yet they can never notate more than a fraction of 
the expression latent in the music and inviting the performer's intuitive 
response. As in other directions, the baroque musicians preferred to 
trust to the performer's instincts not as little but as much as possible, 
and we must always keep the principle in sight: there are fewer expres- 
sion marks in early music, but there was not less expression. 

(413) Giulio Caccini, Nuove Musiche, Florence, 1602, Introd., transl. 
John Playford, Introduction, London, 1654, ed. of 1674, p. 40: 

'In Encreasing and Abating the Voice, and in Exclamations, is the 
foundation of Passion. ... [p. 46:] We see how necessary a certain 
judgement is for a Musician. ... [p. 47:] But because there are many 
things which are used in a good fashion of Singing ... we use to say of 
a man that he sings with much Grace, or little Grace.' 



(414) Francois Couperin, UArt de Toucher le Clavecin, Paris, 1716, 
ed. of 1717, preface: 

'Experience has taught me that powerful hands capable of performing 
whatever is most rapid and light are not always those which show to best 
advantage in tender and expressive pieces, and I declare in all good faith 
that I am more pleased with what moves me than with what astonishes 

(415) Francesco Geminiani, Treatise of Good Taste, London, 1749, 
p. 4: 

'And I would besides advise, as well the Composer as the Performer, 
who is ambitious to inspire his Audience to be first inspired himself, 
which he cannot fail to be if he chuses a Work of Genius, if he makes 
himself Thoroughly acquainted with all its Beauties; and if while his 
Imagination is warm and glowing he pours the same exalted Spirit into 
his own Performance/ 

(416) Charles Avison, Essay on Musical Expression, London, 1752, 
p. 89: 

'For, as Musicdl Expression in the Composer, is succeeding in the 
Attempt to express some particular Passion; so in the Performer, it is to do 
a Composition Justice, by playing it in a Taste and Stile so exactly corre- 
sponding with the Intention of the Composer, as to preserve and illustrate 
all the Beauties of his Work/ 

(417) Joachim Quantz, Essay, Berlin, 1752, XI, 5: 

'The good effect of the music depends almost as much on the perfor- 
mers as on the composers. 

[9:] "Almost every musician has a different expression from that of 
others. It is not always the different teaching that they have received 
which causes this variety; the difference of temperament and character 
also contribute. 

[15:] 'The performer of a piece ought to seek out and arouse in himself 
not only the main feeling but the others which occur. And since in most 
pieces there is a perpetual interchange of feelings; the performer should 
know how to judge which is the feeling in each idea, and to govern his 
expression accordingly/ 

(418) C. P. E. Bach, Essay, Berlin, 1753, III, 13: 

'A musician cannot move others unless he too is moved. He has to feel 
in himself all the feelings he hopes to rouse in his hearers, for it is the 
showing of his own emotion which calls up a similar emotion in the 





A practical point of some importance to decide before the performance 
begins is what repeats are to be made, and in the case of suites, which 
movements are to be included. As we should expect under the free 
conditions of baroque performance, both these decisions are within the 
option of the performer, though conventions existed which will influence 
his choice. 


At the beginning of the baroque period, the dots on one or both 
sides of a bar-line, single or double, which now indicate a repeat, had 
not yet acquired this meaning. They were decorative in origin: we find 
them between double bar-lines or beside them, often more numerous 
than our customary pair, and in a variety of pleasing patterns. But 
standing, as they did, at the division between sections which in many 
instances were regularly repeated as a matter of convention, they took 
on the function of repeat marks by gradual stages, and by the latter 
part of the seventeenth century they were established more or less in 
their normal modern use. 

(419) Charles Butler, Principles of Musik, London, 1636, p. 37: 

*A Repete is either of the same notes and ditty words together, having 
this mark J; or of ditty with other Notes having this mark; :||:, or 
this, ij : before which the first word of the Repeated ditty is commonly 
placed under his note or notes: or of a whole Strain [section]; having at 
the ende thereof 2 prickt Bars bars with dots, through all the Rules 

In vocal part-books of this period we find Butler's signs, including 
the double bar with dots to either side, between staves, in the senses he 

(420) John Playford, Introduction to the Skill of Mustek, London, 
1654, ed. of 1674, p. 35: 

*2. Bars are of two sorts, single and double. The single Bars serve to 
divide die Time, according to the Measure of the Semibreve: the double 



Bars are set to divide the several Strains or Stanzas of the Songs or Lessons 
[pieces], and are thus made, 

C 3. A Repeat is thus marked S, and is used to signifie that such a part of a 
Song or Lesson must be Played or Sung over again from that Note over 
which it is placed/ 

Here the sign which looks most like our modern repeat sign has no 
such meaning, and a quite different sign is used for the repeat. Other 
signs of this period which may bear the meaning of repeat signs are 
various clusters of dots such as .". and :: together with a few other 
patterns; but these may also be the equivalent of ^ in its modern sense 
of a pause (not with its common baroque implication of an ornamental 
cadenza). We also find *g. and allied forms, both in the modern meaning 
of the 'segno' (sign) from which a repeat (usually less than a full section 
in length) is to be made; and as the equivalent of o meaning a pause 
with or without ornamentation. 

Georg Muff at (Florilegium /, Augsburg, 1695, preface) gives i - 1 
i - 1 for 'first-time' bars. 

The next quotation states clearly what most authorities evidently 
regarded as too obvious and elementary to need stating at all: i.e. that 
in music falling into regular sections, the double bar itself implies a 
repeat even in the absence of dots before or after it. This confirms what 
has been said above, that it was not originally the dots which indicated 
a repeat, but the double bar or more accurately, the sectional con- 
struction itself which the double bars merely helped to make clear, 
while the dots, in turn, merely ornamented the double bar. 

(421) Benjamin Hely, The Compkat Violist, London [c. 1700], p. 3: 

'When you meet with double Bars these show that you are at ye end 
of the part or Strain, which is to be play'd twice and then to proceed/ 

Brossard (Dictionaire de musique, Paris, 1703, s.v. 'Ripresa'), gives a 
variety of repeat marks recognisably identifiable with the modern sign, 
and also two signs not unlike Playford's repeat sign, but in the special 
sense of the French Petite Reprise. This occurs over the last few bars 
of a section which has already been repeated in its entirety (Grande 
Reprise), when these few bars only are given an additional repeat 
conventionally taken very softly as an echo effect. 

Ex. 189. Sebastien de Brossard, loc. cit., marks (a) for Grande 
Reprise and (b) for Petite Reprise: 



Monteclair (Petite Methode, Paris, c. 1730, p. 36) the following 
for 'first-time* and 'second-time' bars. 

Ex. 190: Michel de Monteclair 9 he. cit. 9 repeat sign and continu- 

(422) Joachim Quantz, Essay, Berlin, 1752, V, 27: 

[Ex. 191 (a)] 'means that the piece has in truth two parts, and that the 
first ought to be repeated; but that the first should not be repeated, until 
after having played all the piece from one end to the other. For then one 
repeats the first part up to the double bar; or what comes to the same 
thing, up to the note which precedes it, and on which one finds an arc with 
a dot [i.e. ^]. 

[Ex. 191 (b) means] 'that it is necessary to play twice the notes which 
follow up to another Bar which has the same dots before it [often Bis is 

[Ex. 191 (c) means] 'that the piece has two parts, and that each part 
must be played twice. But when at the end there are one or two arcs with 
two dots [Ex. 191 (d)] that signifies that the piece finishes at this point/ 

Ex. 191. Quantz, lac. tit., repeat signs, etc.: 

(a) (b) (c) (d) 

The first paragraph of the above quotation describes the normal da 
capo procedure. The word bis in the second paragraph is the Latin for 
'twice'. There is nothing in Quantz' account which is any longer 
inconsistent with modern practice and notation. 


It was standard practice to repeat the sections of dances, whether so 
indicated or not. 

(423) Thomas Morley, Plaine and Easie Introduction, London, 1597, 
p. 181: 

'A pavane . . . most commonly made of three strains, whereof every 
straine is plaid or sung twice.' 

This convention appears to have been universal in renaissance court 
dancing, and to have continued throughout the baroque period. The 
convenience of a regular understanding on the point is obvious. The 
musicians know what to do without being told; the dancers arrange 



their figures on the same assumption. But in addition there is a natural 
symmetry in the regular repeats which is as satisfying to the ear as the 
symmetry of the dance figures is to the eye. 

In accompanying renaissance and baroque dances, the repeats must 
be correctly made. In playing such music purely for its own sake, the 
same course should be followed. There is no good artistic reason for 
departing from it. 

See also (421) in 2 above. 


Music composed in dance form, but meant to be enjoyed for its own 
sake and not primarily, if at all, for dancing, may be less symmetrical, 
in which case the repeats may not so obviously impose themselves as 
they do where the symmetry of the music conditions our minds to 
expect a corresponding symmetry in the repeats. But even in long suites 
it is usually better to make every repeat which the structure of the 
music itself suggests. The sectional form is part of the effect, and the 
repeats which belong to it should be treated as emphasising the sectional 
form. It is usually better to cut movements than to cut repeats, if the 
total length is felt to be excessive. On the other hand, in many cases the 
first section is considerably shorter than the second section: it may then 
work very well to repeat the first but not the second section. 


We have seen in Ch. X, 3 (a) above that more or less impromptu 
variations were especially favoured at the repeat of a section. It is almost 
certain that the ornamented versions of the sarabandes in J. S. Bach's 
English Suites were meant to be used as varied repeats rather than in 
their entirety after the plain versions. 


Suites, sets of variations and even sonatas, were not necessarily 
meant to be performed complete: many 'suites', indeed, are mere 
convenient collections of pieces in the same key. 

(424) Francois Couperin, Concerts Royaux, Paris, 1722, preface: 

'I have composed them [at different times] for the little chamber 
concerts where Louis Quatorze made me come every Sunday in the year 
[but for publication] I have arranged them by keys/ 

(425) Girolamo Frescobaldi, Toccatas, Rome, 1614, preface, 2: 

'In the Toccatas, I have tried not only to provide plenty of divisions and 
expressive variations, but also to arrange the different sections so that they 
can be performed separately from one another, in such a way that the 



performer, without being compelled to perform them complete, can 
stop wherever he prefers. 

[9 :] *The different sections of the PassacaiUes can be performed separ- 
ately, at will, provided you take care to join the different movements 
together; and this is also the case with the Chaconnes/ 

(426) Francesco Maria Veracini, Senate Accademiche, London and 
Florence [1744], Intentions of the author*: 

'Each of these 12 sonatas consists of 4 or 5 movements: take notice, that 
this is done for the enrichment, and adornment of the work, and for the 
greater entertainment of lovers and connoisseurs of music. Nevertheless 
2 or 3 movements chosen at pleasure from each sonata will be sufficient 
to make up a sonata of satisfactory proportions. 9 

To insist on every movement of a lengthy suite or every variation of 
an interminable set is to show a misplaced respect for the composer 
which would have seemed quite meaningless under baroque conditions. 
The gain in force and compactness from leaving out the least interesting 
matter, or indeed simply from reducing the length irrespective of the 
degree of interest, may make all the difference between losing the 
attention of the audience and holding it. 





Tempo in Early Music 


To set a good tempo ; to maintain it, flexibly, yet so that the piece ends 
at the same tempo at which it started; to remember this tempo so as to 
be able to set it again, within a reasonable margin, at the next per- 
formance: these are some of the hardest things in music. 

Notation is at its least helpful here. This is not because tempo cannot 
be recorded in writing, but because, in practice, it is not an absolute 
but a relative quantity. 

The familiar story of Beethoven's irascible inability to believe his 
own previous metronome markings illustrates very well the fallacious- 
ness of assuming that a good tempo at one performance is a good 
tempo for every performance. It is not so ; there are too many variables 
which affect the case. 

Some of these variables are physical. A room, hall or church with 
resonant acoustics imposes a slower tempo than one with little echo. 
Large choirs and orchestras make for a slower tempo than small forces 
in the same music. 

The most important variables, however, are the temperament and the 
passing mood of the performer. Fine music has depths and shades of 
meaning which cannot all be fully brought out in the same performance. 
We can make the most of its brilliant side, of its tender side and so 
forth, but not all at the same time. Not only may different performers 
find different affinities in the same music; the same performer may do 
so at different times. And one of the main changes involved in such 
changes of interpretation is a change of tempo as Beethoven dis- 

We have often the impression that a tempo is so exactly right that 
the least change could only be for the worse. But this is really a com- 
pressed statement. What we mean is that the interpretation pleases us 



so well that none could please us better; and that for this interpretation 
the tempo is exactly right. 

A fast movement taken steadily may gain a breadth of interpretation 
for which no mere speed can compensate; it may even sound more 
brilliant. A slow movement may gain more poignancy from being kept 
well in motion than from being dragged under the illusion of making 
the most of it. No slow movement should make us think of it as slow, 
nor fast movement as fast. A certain pulse should keep the slow move- 
ment alive; a certain ampleness should keep the fast movement from 
sounding hurried. Yet if these conditions can be met, some very slow 
speeds can be made to pay dividends by sheer intensity of feeling, and 
some very fast ones by breath-taking poise and rhythm. 

Some movements allow a much narrower margin of tempo than 
others. Perhaps there is always just one interpretation, and therefore 
just one tempo, which most musicians will find more convincing than 
any other; or perhaps interpretation is always relative. In either case, 
the only way of finding the tempo is by responding to the music itself, 
with a sensitiveness not given to every musician alike nor to any 
metronome at all. 


(a) Attempts have been made to establish an agreed standard of 
tempo by relating, for example, the crotchet to the pulse-rate of the 
human heart (Mersenne, Harmonie Universelle, Paris, 1636-7; Simpson, 
Compendium, London, 1665; Quantz, Essay, Berlin, 1752), to the tick 
of a watch (Simpson, loc. cit.), to the swing of a pendulum (Galileo, 
16th cent. ; Mace, Mustek's Monument, London, 1676; Loulie, Prindpes 
de musique, Paris, 1696; Tans' ur, New Musical Grammar, London, 
1746), or merely to time in the abstract (|TAbb6 Laugier] Apohgie de 
la Musique Franfoise [Paris], 1754). The Maelzel metronome (a com- 
pound pendulum of manageable size, actually invented or developed by 
Winkel) was perfected by 1816, since when we have possessed a reliable 
and agreed means of notating tempo. But for the reasons given above 
it is of limited value in practice, and can be misleading if this limitation 
is not understood. 

(427) Anton Bemetzrieder, Lepons de Clavecin, Paris, 1771, p. 68: 

'Taste is the true metronome.' 

For the same reasons, time-words and time-signatures, besides being 
much vaguer, can be equally misleading. The evidence concerning these 
in the contemporary authorities is very extensive, and we cannot afford 
to ignore it; but it is also very contradictory and on the whole very 
out of touch with practical realities. Indeed, having aired the theories 



of tempo which, seem to have been conventionally expected of them, 
most authorities warn us that these have not much/esemblance to the 
practice of musicians, and tell us to judge our tempos from the music 
itself. It is most important to remember this whenever we are puzzled 
by a time-word or a time-signature, or any other problem of early 
tempo. We must be guided by the music and not by the notation. 

(b) For a seventeenth-century warning of the perpetual tendency of 
notated time-values to slow down, thus seriously misleading later users, 
see (564) in Ch, XLII, 2 below. Many modern editors, including some 
of the best, prefer to halve or even to quarter the note-values of much 
early music in order to protect the performer from playing what appear 
to be long notes far too slowly. When the discrepancy between appear- 
ance and reality is very wide, this course is desirable in spite of the fact 
that it can also be misleading, chiefly on the natural pulse of the music. 
Performers are nearly always to some extent hypnotised by what they 
see in notation. Ideally they should be able to grasp that a minim in 
renaissance music was still very nearly what its name (minima nota, 
'the smallest note') describes and play accordingly. Actually the mere 
sight of breves, semibreves and minims does slow most people down to 
a funeral pace, which for a gay dance or part-song of the sixteenth 
century is very damaging. But where, as in much seventeenth-century 
music, the discrepancy between the apparent and the real note-values 
is not very wide, to halve the original note-values may easily result in 
too fast a speed, and it may be better not to make this change. 


The following is interesting, since both authors were pupils of J. S. 
Bach, and one was his most gifted son. 

(428) C. P. E. Bach and Johann Friedrich Agricola, joint article in 
Lorenz MMer's Musikalische Bibliothek, Leipzig, 1739-54, V, iii: 

'[J. S. Bach] was very accurate in Ms conducting and very sure of Bis 
tempo, which he usually made very lively/ 

That does not, of course, tell us how lively, or when. It does not 
justify empty speed; but it does tell us not to be afraid of a virtuoso 
tempo in movements which suggest it. We may set against it J. S. Bach's 
mention of a 'slow time' in Italy and Germany in the only document 
directly recording his own teaching on tempo which seems both brief 
and perfunctory, as if he had decided that there is not much to be said 
about tempo ; the only way to learn it is through your ears. 

(429) J. S. Bach, ? lecture notes taken at dictation in the hand- 
writing of his pupil, Johann Peter Kellner, in a MS. entitled Instructions 
and rules for the playing of a Thorough-Bass ... by Master John 



Sebastian Bach', and most of it copied, presumably approval, 

from F. E. Niedt's Musicalische (Hamburg, 1700); printed 

complete In PMlipp Spitta's J. S. (Leipzig, 1873-80, transl. Clara 
Bell and J. A. Fuller-Maitland, London, 1884-85): 


'Of this much need not here be said for it is presupposed that a person 
wishing to learn figured-bass will not only have learnt the notes but also 
the intervals before doing so, whether by previous practice of music or 
from some other cause, and also the differences of time. For no one can 
inculcate a knowledge of time all at once. This must, however, be noticed, 
that in the present day one single kind of time is indicated in two ways, 
thus: C2, the second way being used by the French in pieces that are 
to be played quickly or briskly, and the Germans adopting it from 
the French. But the Germans and Italians abide for the most part by the 
first method, and adopt a slow time. If the piece is to be played fast the 
composer expressly adds Allegro or Presto to it; if slowly, the pace is 
indicated by the word Adagio or Lento.* 

(430) Joachim Quantz, Essay, Berlin, 1752, XVII, vii, 52: 

'The sense of the words should be taken into account, the movement 
of the notes, especially the fastest, and in quick arias the skill and the 
voice of the singer. A singer who pushes his quick passages with his chest 
can scarcely take them as fast as one who only sings them from Ms throat, 
though the first will always excel the second in clarity, particularly in a 
large place. . . . 

[53:] It is the same with church music as it is with arias; except that 
both expression and tempo should be more restrained than in opera, to 
show respect for the sacredness of the place.* 





Time-words are notoriously vague. They often relate strictly to mood, 
not to tempo: e.g. largo (broadly), grave (gravely), adagio (at ease), 
maestoso (with majesty), allegro (cheerfully), etc. Tempo is a function 
of mood, rather than the other way about. 

(431) A. Malcolm, Treatise of Mustek, Edinburgh, 1721, p. 394: 

'Time ... is a various and undetermined thing . . . tho' the same 
Notes of Time are of the same Measure in any one Piece, yet in different 
Pieces they differ very much, and the Differences are in general marked 
by the Words slow, brisk, swift, etc., written at the Beginning; but these are 
still uncertain Measures, since there are different Degrees of slow and 
swift; and indeed the true Determination of them must be learnt by 
Experience from the Practice of Musicians.' 

(432) Joachim Quantz, Essay, Berlin, 1752, XI, 15: 

'There are indeed various degrees of liveliness and sadness, for example 
in a furious mood an expression is needed which has more fire than in a 
witty piece, although both the one and the other ought to be lively. 

'[16: The prevailing mood can be judged (1) by the mode major or 
minor; (2) by the intervals close or leaping; (3) by the discords mild or 
harsh; (4) by] the word found at the beginning of the piece, such as: 
Allegro, Allegro non tanto . . . etc. All these words, provided they are 
not put without reason, demand each the use of a particular expression 
and moreover, as we have already said, since each piece of the kinds 
described can include a mixture of pathetic, tender, gay, sublime and 
witty thoughts, each must therefore be given, so to speak, another mood, 
to be sometimes sad, sometimes gay, sometimes serious, etc., the which 
interplay is absolutely necessary in music. 

[17:] 'Each [performer] ought in this respect to conform to his tempera- 
ment and know how to guide it suitably. 

[XII, 2:] 'Since many composers put these words above mentioned 
more from habit than to characterise well the true movement of the 
pieces, and to assist the knowledge of their true time for those who render 
them; there are many cases where they cannot be used for guidance, and 
where it is necessary to divine the intention of the composer more from 
the content of the piece than from the word which is found at the head to 
indicate its movement. 



[11:] "Whatever speed an Allegro demands, it ought never to depart 
from a controlled and reasonable movement. For everything that is 
hurried, causes the ear more pain than satisfaction. The object most 
always be the feeling which is to be expressed, never only to play fast. 

[24:] 'The mood often changes as much in Allegro as in Adagio. 

'[XVII, i, 6: it is for the composer] to leave it to the player of the 
principal part to take the time which he judges suitable.' 

(433) C. P. E. Bach, Essay, Berlin, 1753, III, 10: 

'The tempo of a piece, which is usually indicated by a variety of familiar 
Italian terms, is derived from its general mood together with the fastest 
notes and passages which it includes. Proper attention to these considera- 
tions will prevent an allegro from being hurried and an adagio from 
being dragged.' 

(434) Leopold Mozart, Violinschule, Augsburg, 1756, 1, ii, 7: 

'It is true that special words are written at the start of each piece which 
are supposed to give its character, such as Allegro (lively), Adagio (slow) 
and so forth. But both slow and fast have their gradations, and even if the 
composer tries to describe more plainly the speed required by adding 
further adjectives and other words, it still remains impossible for him to 
describe accurately the speeds he wants for the performance of the music. 
Thus it must be inferred from the music itself, and this is what infallibly 
shows the true quality of a musician. Every melodic piece includes one 
phrase at least from which the variety of tempo needed by the music can 
be clearly recognised. This phrase, if other considerations are taken into 
account, often compels one into its own natural speed. Bear this in mind, 
but also realise that for this sort of intuition long experience and fine 
sensibility are required. Who will contradict me if I regard this as among 
the highest accomplishments in the art of music?' 

The above quotations confirm that we should let the music suggest 
its own tempo even if this appears to be contradictecTby the time-word. 

They also warn us against one of the commonest mistakes in the 
interpretation of early music: taking slow movements too slow and fast 
movements too fast. 


The few time-words which appear in early baroque music were more 
clearly confined to tempo, and less indicative of mood, than later 
became the case. 

We find tardo (slow), presto (fast) in Biagio Marini (Sonatas, etc., 
Venice, 1626) and others; long', 'slow', 'away' in Orlando Gibbons 
(fantasies, Christ Church, Oxford, MSS. 732-5, early 17th cent.) and 
subsequent English sources. The number of such words increased later 




in the seventeenth century, until our familiar vocabulary became 
developed particularly in Italian and French. 


Considerable discrepancies will be noticed in the order of speed 
given in the following lists. We find, too, inconsistent time-words in 
different contemporary copies of the same piece, and even in different 
part-books of the same edition. This was not because tempo was 
regarded as unimportant, but because, as usual, the performer was 
expected to use his judgment with a minimum of textual help. 

The entries from Brossard and Grassineau occur, of course, in their 
alphabetical positions in the dictionaries concerned. For ease of com- 
parison, they are shown here, like the others, in ascending order of 
speed so far as this is shown by the descriptions. Leopold Mozart's list, 
with its valuable hints on style, is also shown here in ascending order 
of speed, for the same reason; but unlike most such lists, his is shown 
by the original edition in descending order of speed. The reader who 
wishes to read this list in its original order should therefore start with 
the last entry as shown here, and work his way backwards to the 
beginning. The same applies to Rellstab's list. 

(435) Henry Purcell, Sonnatcts of III Parts, London, 1683, preface: 

'Adagio and Grave . . . import nothing but a very slow movement: 
Presto Largo, Poco Largo, or Largo by it Self, a middle movement/ 

(436) Sebastien de Brossard, Dictionaire de musique, Paris, 1703: 

'LARGO . . . VERY SLOW, as if enlarging the measure and making the 
main beats often unequal, etc. 

'ADAGIO ADAGIO means very slow. 

'ADAGIO . . . COMFORTABLY, at your ease, without pressing on, thus almost 
always slow and dragging the speed a little. 

'LENTO means SLOWLY, heavily, not at all lively or animated. 

'AFFETTO, or con Affetto. This is the same as Affettuosb or Affettuosamente, 
which means FEELINGLY tenderly and thus nearly always Slow. 

'ANDANTE . . . to stroll with even steps, means above all for Basso Con- 
tinues, that all the Notes must be made equal, and the Sounds well 

'ALLEGRETTO diminutive of Allegro, means RATHER .GAILY, but with a 
gracious, pretty, blithe gaiety. 

'ALLEGRO . . . always GAY, and decidedly lively, very often quick and 
light; but also at times with a moderate speed, jttgay, and lively. 

'ALLEGRO ALLEGRO marks an intensification of gaiety or of liveliness, etc. 

'PRESTO means, FAST. That is to say the speed must be pressed on, by 
making die beats very short. 


'PRESTO PRESTO or Prestissimo. Means, very quick, 9 

(437) Alexander Malcolm, Treatise of Musick, Edinburgh, 1721, p. 

'Grave, adagio, largo, vivace, allegro, presto, and sometimes prestissimo. 
The first expresses the slowest Movement, and the rest gradually quicker; 
but indeed they leave it altogether to Practice to determine the precise 

(438) James Grassineau, Musical Dictionary 9 London, 1740 (the 
following entries are not, like most of this work, taken from Brossard's 
Dictionaire, Paris, 1703, for which see above): 

'ADAGIO . . . expresses a slow time, slowest of any except grave 
[British Museum copy corrected in ink by Dr. Bumey to "the slowest of 

'ALLEGRO . . . brisk, lively, gay and pleasant ... yet without pre- 

'LARGO, a slow movement, i.e. one degree quicker than grave, and 
two than adagio. 

'LENTE, or LENTAMENTE, signifies a slow movement, much the same 
as largo. 

'LENTO, the same as lente. 

'CANZONE, in general signifies a song, wherein some little fugues 
are introduced. ... If played in any part of a Sonata, it signifies much 
the same as Allegro, and only denotes that the part to which it is pre- 
fixed is to play or sing in a brisk and lively manner. 

'PRESTO, fast or quick, gayly yet not with rapidity. PRESTISSIMO, is 
extreamly quick, hastily, with fury/ 

(439) Leopold Mozart, Violinschule, Augsburg, 1756, 1, iii, 27: 

'GRAVE, sadly and earnestly, hence very slow indeed. 

'LARGO, a still slower tempo [than Adagio Pesante], must be played 
with long bow-strokes and with much tranquillity. 

'ADAGIO PESANTE, a sad Adagio, must be played somewhat slower [than 
Adagio], and dragging. 

'ADAGIO, slow. 

'ANDANTE, going (gehend). This word itself tells us that one must 
give the piece its natural gait (Gang) ; especially when MA UN POCO 


'MAESTOSO, with majesty, deliberately, not hurried. 

'SOSTENUTO means sustained, or rather held back and the melody 
not exaggerated. One must also make use of earnest, long and sustained 
bow-strokes in such cases, and hold on to the melody in a very legato 
manner (den Gesang wohl aneinanderhangen). 

'TEMPO COMMODO and TEMPO GIUSTO lead us in the same way [as 



Moderate] back to the piece itself. They tell us that we must play 
the piece neither too fast nor too slow, but in a suitable, convenient and 
natural tempo. We must therefore look for the true pace of such a piece 
in the piece itself. 

'MODERATO, moderately, temperately; not too quick and not too 
slow. This again the piece itself must show us, we must become sensible 
of the moderation during the course of it. 

'VIVACE means lively, and SPIRITOSO says that we should play 
with intelligence and spirit, and ANIMOSO is almost the same. All 
three kinds stand midway between fast and slow, as to which the piece of 
music bearing these words must itself show us more. 

'ALLEGRETTO is somewhat slower than ALLEGRO, has usually something 
agreeable, something pretty and playful, and has much in common 
with ANDANTE. It must therefore be performed prettily, lightly and 
playfully: which prettiness and playfulness can be as clearly indicated, 
in this as in other tempos, by the word GUSTOSO. 

say that one must not exaggerate the speed. For this a lighter and more 
lively, yet at the same time somewhat more earnest and broader bowing 
is required than in a quicker tempo. 

'ALLEGRO which indeed shows a gay, but not hurried tempo; 
especially when it is moderated by adjectives and adverbs [as above]. 

'MOLTO ALLEGRO is somewhat less than ALLEGRO ASSAI, but is still faster 
than [the foregoing]. 

'PRESTO means fast, and ALLEGRO ASSAI is very little different. 

'PRESTISSIMO shows the fastest tempo, and PRESTO ASSAI is almost the 
same. For this rapid time a light and somewhat shorter stroke is needed.' 

(440) J.-B. Cartier, UArt du Violon, Paris [1798], list of speeds in 
ascending order: 

'Largo, Larghetto, Adagio, Grave, Andante, Andantino, Grazioso, 
Affectuoso, Amoroso, Moderato, Tempo Giusto, Maestoso, Allegro, 
Allegretto, Allegro Molto, Allegro con moto, Allegro Agitato, Allegro 
Spiritoso, Allegro assai, Vivace, Presto, Presto assai, Prestissimo.' 

The following requires special attention. 

(441) Sebastien de Brossard, Dictionaire de Musique, Paris, 1703: 

'ASSAI . . . which the Italians often join with Allegro, Adagio, Presto, 
etc., means, according to some, MUCH; and according to others that 
neither the measure nor the tempo should be carried to excess, but that a 
judicious mean of slowness, and of rapidity should be preserved, according 
to the different impressions it is necessary to convey.' 


Quantz (Essay, Berlin, 1752, Ch. XVII, Sect, vii, 49 ff.) iUustrates his 
scheme for measuring tempo by the human pulse (treated as averaging 



80 per minute) by giving four, later increased to five of time- 

words. These are not actual tempos, but averages which he regards as 
a suitable 'foundation' for teaching purposes. They cover very wide 
extremes of speed, and are equivalent to the following metronome 

M.M. numbers based on Quantz 

I: Allegro assai (including Allegro molto, Presto, etc.) J = 160 
II: Allegro (including Poco Allegro, Vivace, etc.) j = 120 

III: Allegretto (including Allegro ma non tanto, non 

troppo, non presto, moderate, etc.) J = 80 

IV: Adagio cantabile (including Cantabile, Arioso, Lar- 
ghetto, Soave, Dolce, Poco Andante, Affettuoso, 
Pomposo, Maestoso, Alia Siciliana, Adagio 
spiritoso, etc.) J = 40 

V: Adagio assai (including Adagio pesante, Lento, 

Largo assai, Mesto, Grave, etc.) J = 20 

The above values are c in Common time'. 
Double the speed 'in Alia breve time*. 

Quantz (loc. cit.) gives the same theoretical distinction of 'full time 
or alia breve' for triple time, where it corresponds to C or < set before 
numerical triple-time signatures : see Ch. XXXVIII, 5 below. The reality, 
as we shall find, is by no means so simple. 



Dance Tempos 


(a) Dance steps can only be performed correctly within narrow margins 
of speed. It is thus possible to discover the tempo of a dance by actually 
reconstructing it and dancing it. For pioneer work in this direction, see 
my Bibliography under Mabel Dolmetsch and John Guthrie. For a 
more recent contribution, see under Imogen Hoist. 

It must, however, be remembered that both the steps and the figures 
of a dance may have varied widely, sometimes almost unrecognisably, 
at different times and places; and with them, the tempo. 

It must further be remembered that dances which have once left the 
dance-floor and become musical forms in their own right almost 
inevitably undergo some modification, and usually a considerable 
transformation. They tend to slow down as well as growing more 
flexible in rhythm. Their style may be more sophisticated, their figura- 
tion more elaborate, their mood more introspective. A Sarabande by 
J. S. Bach must clearly be interpreted in the light of its own context, 
not as it might be if it were a mere background for dancing. 

(442-3) Ch. Burney, Music in Germany, London, 1773, I, p. 162: 

'[This Polish nobleman] would gladly give me a specimen of the 
[violin] music of his country, as it depended so much on the coup d'archet, 
that seeing it on paper, without hearing it performed, would afford but 
a very imperfect idea of it. 

'[The Pole added that] the kind of music which we call Polonaise, is 
played quicker for dancing than at other times/ 

(b) The fluctuations of the commoner dances and dance-forms can 
be followed to some extent in the following quotations, which are 
taken from Thomas Morley (Plaine and Easie Introduction, London, 
1597, ed. of 1952, pp. 296 ff; Thomas Mace, Mustek's Monument, 
London, 1676, pp. 129 ff.); James Talbot (manuscript notes, Christ 
Church, Oxford, MS. 1187, c. 1690); Ch. Masson (Nouveau Traite, 2nd 
ed., Paris, 1701, pp. Iff.); Sebastien de Brossard (Dictionaire de Mu- 
sique, Paris, 1703); Johann Gottfried Walther (Musicalisches Lexicon, 
Leipzig, 1732); Jean- Jacques Rousseau (in the Encyclopedia of Diderot 
and d'Alembert, Paris, 1751-65, s.v. 'Minuet', repeated almost verbatim 


In Ms own de Paris, 1768, s.i\ *Minuef); 

Quantz (Essay, Berlin, 1752, XVII, vii, 58). 
See also Ch. XXXVIII below for further to 


Two main forms, an earlier and more rapid, and a later more 
steady (as in J. S. Bach). 

(444) Morley, he. dL, 1597: 

4 The Alman Is a more heavie daunce than [the Galliard] (fitlie repre- 
senting the nature of the [German] people, whose name it carrieth), so 
that no extraordinarie motions are used in daunclng of it/ 

(445) Mace, he. cit., 1676: 

'Allmaines, are Lessons very Ayrey, and Lively; and Generally of Two 

Strains, of the Common, or Plain-Time J 

(446) Talbot, he. at, c. 1690: 

'Almain so called from Almaine or Germany from whence it came, 
this is sett in Common the same time as the Pavan, but its movement is 
somewhat quicker and more Airy.' 

(447) Brossard, op. cit. 9 1703, 3rd ed. [1707?]: 

'Allemanda grave symphony, usually in two time, often in four.' 

(448) Walther, he. cit. 9 1732: 

'Allemanda ... is a serious and dignified movement and should be so 


Not a dance-form, but very like one in seventeenth-century England: 
also developed to a serious polyphonic movement associated with the 
fantasy for viols, and in this shape often profoundly beautiful, yet 
always with an 'ayery', i.e. a lyrical quality. 

(449) Mace, he. cit. 9 1676: 

'Ayres, are, or should be, of the same Time [as Allmaines] . . . only they 
differ from Allmaines, by being commonly Shorter, and of a more Quick, 
and Nimble Performance [but some Ayres are slow]-' 


A generic name for a large species of renaissance round dances, with 
stamping of the feet. 

(450) Morley, he. cit. 9 1597: 



"Like unto [the Alman] is the French Bransle (which they call branle 
simple) which goes somewhat rounder in time than this, otherwise the 
measure Is all one. 

The bransle de Poictou or braisle double is more quick in time.' 


A dance of which the movement resembles the gavotte, but with a 
pronounced pulse of two in the bar, whereas the gavotte has four. 

(451) Talbot, he. cit., c. 1690: 

*Bor<e [one of five] French Measures of a very quick and rapid Move- 

(452) Masson, he. tit., 1701 : 

'Bourse and Rigaudon, quicker [than Gavotte).' 

(453) Quantz, he. cit., 1752: 

4 Bourr6e and Rigaudon played gayly and with a short and light bow- 
stroke. Each bar has one beat of the [human] pulse/ 


A quick dance with a forthright dotted rhythm like an English jig. 

(454) Talbot, be. cit., c. 1690: 

'[Same as for Boree, above] very quick and rapid movement.' 

(455) Masson, he. cit., 1701 : 
'Canaries a little quicker than Gigue/ 

(456) Quantz, loc. cit., 1752: 

'The Gigue and the Canarie have the same movement. If written in 
6-8 time, each bar has one [human] pulse beat. ... In the Canarie, 
which consists always of dotted notes, the bowing is short and sharp/ 


A dance with both duple-time and triple-time variants, of which the 
former did not persist. 

Of the triple-time variants, there are two sub-variants whose 
differences are important. 

The Italian coranto is quick, and normally straightforward in 
rhythm: literally 'running'. 

The French courante has a much more sophisticated rhythm, with a 
very characteristic alternation of simple with compound triple time 



(mainly 3-2 and 6-4). To make this effective, a 

and a pointed style are needed: no longer "running' in the 

(457) Morley, loc. cit., 1597: 

'Like unto [the bransle double] be the voltes and couraetes which 
both of a measure, are notwithstanding daunced after sundrie 
. . . the courante travising [traversing, i.e. crossing] and running [whence 
the name]/ 

(458) Mace, loc. cit. 9 1676: 

t Corantoes 9 are Lessons of a Shorter Cut [than Galliards], and of a Quicker 
Triple-Time; commonly of 2 Strains, and full of Sprightfiilness, and 
Vigour, Lively, Brisk, and Cheerful! 

(459) Masson, loc. at, 1701 : 
'Courante [is taken] steadily (gravement). 9 

(460) Walther, loc. at., 1732: 

'The measure of the Courante or rather the rhythm of the Courante 
dance is the most solemn of any [an exaggeration; but the contrast of this 
slow French courante with the quick Italian coranto is instructive].* 

(461) Quantz, loc. tit,, 1752: 

'The Courante [is] performed with majesty, and the bow is detached at 
each crotchet, whether there is a dot or not. Count one beat of the [human] 
pulse for each crotchet/ 


Of the various definitions by which this celebrated variation form 
has been differentiated from the Passacaille, whether in contemporary 
or modern authors, none can be consistently maintained, and indeed 
works are found which bear one heading in one source and the other 
in another source. A lilting triple time is common to both (except for 
a very rare and untypical duple-time variant); but the actual tempo 
required not only differs in different specimens, but in some of the 
more freely-worked and elaborate specimens needs to be varied in 
course of the work. See below under Passacaille. 

(462) Masson, loc. cit., 1701 : 

'The Chaconne [is taken] lightly (legerement)/ 

(463) Quantz, loc. cit., 1752: 

'A Chaconne is also [like the Sarabande] performed with majesty. One 
beat of the [human] pulse goes for two crotchets/ 




Besides its more general meaning of introduction, the word was used 
as follows. 

(464) Walther, he. cit. 9 1732: 

4 The Entree is a serious melody with two strains, for instruments 
only . . . taken with two beats in the bar and used for dance or theatre 

(465) Quantz, he. cit., 1752: 
'{Same as for Courante above.]' 


(466) Quantz, loc. cit., 1752: 

*A Furie is played with great fire. Count one beat of the [human] pulse 

for two crotchets, whether in Common or in Triple Time/ 

(467) John Hoyle, Dictionary of Music, London, 1770: 

'FURIA, or CON FUJUA, is with fury and violence, and is to be 
understood not so much with respect to the loudness of the sound, as to 
the quickness of the time or movement/ 


In the ballroom, the Galliard was paired with the Pavan, and it 
often is so paired as a musical form. 

The Pavan was the main ceremonial entry of the ball during the 
sixteenth century. It is a processional dance in which the entire company 
took part, with dignified, slightly swaying steps in a stately rhythm of 
four in a bar. But the Galliard which followed was intended only, as a 
display of agility and skill, for the younger and more energetic members 
of the party. Its steps are vigorous, and may be elaborated almost to 
the virtuosity of a ballet solo (including entrechats). It is therefore 
commonly described as more brisk and lively than the Pavan. 

This description has given rise to a misconception. Modern writers 
have frequently stated that the tempo of the Galliard is faster than the 
tempo of the Pavan; this is hardly the case. The tempo remains nearly 
the same, each of the four time-units in one bar of the Pavan being 
similar to each of the three time-units in one bar of the Galliard: 
Le. the pulse keeps nearly the same speed as before, in triple time. 
What changes is the dance, not the pulse. The dancer moves faster 
though the music does not. He has now to fit many complex movements 
into the same time which in the Pavan was occupied with one simple 
movement, and if the music were much increased in speed he would 


not be able to perform at all The Galliard, in Is a 

dance with a moderate pulse. 

Even away from the ballroom considerations no 

is intended for a rapid tempo. Indeed, when is no and no 

steps to fit in, there may still be musical variations of such that 

any increase in speed would make them an absurdity if not an 

(468) Morley, loc. cit. 9 1597: 

'After every pavan we usually set a Galliard . . . This is a lighter and 
more stirring kincte of dauncing than the pavane.' 

(469) Mace, he. cit., 1676: 

'Galliards, are Lessons of 2 or 3 Strains, but are performed in a Stow, 
and Large Triple-Time; and (commonly) Grave, and Sober" 

(470) Talbot, foe. cit. 9 c. 1690: 

'Galiard so called from Gallia or France, where it was first used [this 
etymology is incorrect; the derivation is from Fr. gaillard, "a young 
gallant"], a lofty Frolic Movement suitable to the gay temper of the 
Nation, is properly set in a pretty brisk Triple [but see my explanation 

(471) Masson, loc. at., 1701 : 
'Galliards [are taken] lightly (legerement).' 


The pulse of the gavotte is a fairly rapid four in a bar. 

(472) Talbot, loc. eft., c. 1690: 

c [Same as for Boree, above] very quick and rapid movement.' 

(473) Masson, loc. cit., 1701 : 
'Gavottes [are taken] lightly (legerement).' 

(474) Quantz, loc. cit., 1762: 

'A Gavotte is almost equal to a Rigaudon; it has, however, a steadier 


Another dance of which the musical pulse is steadier than the steps. 

(475) Talbot, loc. cit., c. 1690: 

'Horn-pipe, [one of] two very nimble Movements peculiar to the 
English race . . .' 




(476) Talbot, he. a/., c. 1690: 

fi [Same as for Horn-pipe, above.]' 

(477) Walther, he. eft., 1732: 
*GIga ... a rapid English dance.* 

(478) Quantz, he. cit., 1752: 

*The Gigue and the Canarie have the same movement. If written in 
6-8 rime, each bar as one [human] pulse beat. . . . The Gigue is played 

with a short and crisp bow-stroke.* 


(479) Brossard, op. cit., 1703, 3rd ed. [1707?]: 

*Loure . . . ordinarily written under the measure of 6-4 and beaten 

slowly and gravely and marking the first beat of each bar more perceptibly 

than the second/ 

(480) Walther, loc. cit. 9 1732: 

'Loure ... a piece or dance, usually set in 6-4, to be taken in a dignified 
and slow fashion. The first note of each half bar has a dot, which is to 
be well prolonged.' 

(481) Quantz, he. cit., 1752: 
'[Same as for Courante, above.]' 


(482) Talbot, loc. cit., c. 1690: 
'March, a lively martial Movement.' 

(483) Quantz, loc. cit., 1752: 
*A March is played seriously.' 


The Minuet has little in common with the Branle de Poitou, from 
which it has been derived by writers from Praetorius onwards, but a 
great deal in common with the Galliard, of which it is more likely to 
be a development (in the mid-seventeenth century). For no very clear 
reason it is often thought, especially by theatrical producers, to have 
been a vague, mincing, ceremonious and over-refined dance of courtly 



decadence; but none of this is true. The 
and figures, and is by no means lethargic. Its versions 

vigorous than its later versions, but it never the and 

scraping and aimless shifting with which the modern has it 

to caricature it. 

Where two minuets are placed in succession, the direction 4 [First] 
Menuet da capo' Is sometimes found; and this is normally the intention, 
even when the direction does not appear. The second minuet may be 
in the minor of the same key of which the first minuet is in the major; 
and in any case, the da capo is part of the design. 

(484) Talbot, he. cit., c, 1690: 

'[Same as for Bore, above] very quick and rapid movement.' 

(485) Brossard, he. cit. 9 1703: 

'Minuet ... a very lively dance; which comes originally from Poitou 
[but see above]. One ought in imitation of the Italians to use the sign 3-8 
or 6-8 to mark the movement, which is always very gay and very fast; 
but the custom of marking it by a simple 3 or triple crotchet time has 

(486) Masson, he. cit. 9 1701: 
'The Minuet [is] quick.' 

(487) Jean-Jacques Rousseau, he. cit., 1751-65 (but cited from op. 
cit. 9 1768): 

'According to [Brossard] this dance is very gay and its movement is 
very quick. But on the contrary the character of the Minuet is grave and 
a noble simplicity; the movement of it is rather moderate than quick, and 
one might say that the least gay of all the kinds of dance used in our balls 
is the Minuet. It is another matter in the theatre.' 

(488) Quantz, he. cit., 1752: 

'The Menuet is played in such a fashion that it almost carries or lifts the 
dancer up, and one marks the crotchets with a somewhat heavy, yet short, 
stroke of the bow. Count one [human] pulse beat for two crotchets.' 


This dance undoubtedly derives its character from the beautiful 
species of bagpipe known by the same name, and extremely fashionable 
at the French court and in French polite society generally during the 
seventeenth and eighteenth centuries. The most conspicuous feature of 
the Musette as a musical form is its drone harmony. 



(489) Quantz, be. eft., 1752: 

*A Musette Is given a very caressing expression. Count one beat of the 
[human] pulse for each crotchet in 3-4 but for each quaver in 3-8. Some- 
times certain dancers take it into their heads to have it played so fast that 
there is only one beat of the [human] pulse to a whole bar/ 


See above under Cfaaconne; the two were always confused in 
theory and virtually interchangeable in musical practice, and the 
evidence contrasting them is therefore to be viewed with exceptional 

(490) Talbot, loc. cit. 9 "c. 1690: 

*The Passacaille [is taken] gravely (gravement). The Chaconne [is taken] 

lightly (legerement) [but see my remark above].' 

(491) Walther, loc. eft., 1732: 

'Passacaglio . . . actually is a Chaconne. The sole difference lies in its 
being played more slowly . . . and with a less vigorous expression [the 
opposite of Talbot's view!]/ 

(492) Quantz, loc. eft., 1752: 

*A PassecaiUe is equivalent to a Chaconne, but is played a little faster 
[again the opposite of Talbot's view]/ 


As a dance, this is virtually a Minuet taken at a faster speed, as a 
result of which it is easier to dance (because balance is easier). 

(493) Talbot, loc. eft., c. 1690: 

'[Same as for Bor6e, above] very quick and rapid movement/ 

(494) Masson, loc. eft., 1701 : 
'The Passepied [is taken] very quick/ 

(495) Quantz, loc. cit., 1752: 

*A Passepied is played a little more lightly and quickly than a Menuet/ 


See under Galliard, above. The pulse of a Pavan is a very steady 
four in a bar. The dance scarcely survived, at least in fashionable 
circles, into the baroque period, but the form did, and a great many 
attractive specimens appear in the manuscripts of the English chamber 
music of the viols during the first half of the seventeenth century. The 


polyphony of the of Is the 

of the school 

The of the word are very (pavan, 

pavane, efc.), but, in English, the correct pronunciation is pavn. 

The pronunciation pavane is French. 

(496) Morley, loc. at., 1597: 

*A pavane, a kind of staide musicke, ordained for grave dauncing.' 

(497) Mace, loc. cit., 1676: 

'Pavines, are Lessons of 2, 3 or 4 Strains, very Grow, and Sober; Full of 
Art, and Profundity, but seldom us'd, in These our Light Days.* 

(498) Talbot, he. ciL, c. 1690: 

*Pavan ... a grave and stately Movement sett in common time, and 
consisting of 3 strains, this is not now so much in use as formerly/ 


A dance of very characteristic quality with a swinging pulse of two 
in a bar; often the melody is less jerky than a bourrfc and has a longer 

(499) Masson, he. cit. 9 1701 : 

*Bour6e and Rigaudon [are] quicker than [Gavottes and Galliards]/ 

(500) Quantz, he. cit., 1752: 

* A Bourree and a Rigaudon are played gayly and with a short and light 
stroke of the bow. Each bar has one beat of the [human] pulse/ 


The pulse of the baroque Rondeau may be two or three in a bar. 

(501) Quantz, he. cit., 1752: 

C A Rondeau is played with a certain serenity, and one beat of the 
[human] pulse makes almost two crotchets, either in the Allabreve or in 
the 3-4.' 


The original dance has sinuous and complicated movements, and 
had a general reputation for lasciviousness. As a musical form, the 
slow Saraband requires considerable intensity of feeling, often of a 
sensuous variety. The quick Saraband is piquant and virile. J. S. Bach's 
Sarabandes, which are slow, include some of his most impassioned 
harmony combined with a contemplative inwardness which is perhaps 
unique; but some of the French Sarabandes approach them very nearly, 



and in the same mood. The very quick English Sarabands of the mid- 
seventeenth century, with their taut rhythm and nervous energy, are so 
different as to be for all practical purposes another form; they have 
one pulse to each triple-time bar. The following quotations refer to 
Sarabands of different speeds. 

(501 a) Phillips, etal eds., The New World of Words, London, 1658, 
s.v. 'Saraband": 

[Eds. 1-4:] 'Lesson or Air in Musick going with a quick time. 

4 [Fifth Ed. of 1696, partly revised by Purcell, leaves out the tempo 
definition. The B.JVL copy of the 1st ecL, owned by Rd Kendall in 1719, 
has a hand-written correction of "a quick time" to "a slow time" 
particularly interesting evidence of a changing usage.]' 

(502) Mace, foe. cir., 1676: 

'Serabands, are of the Shortest Triple-Time; but are more Toyish, and 
Light, than Corantoes; and commonly of Two Strains. 9 

(503) Talbot, loc. eft, c. 1690: 

'Saraband a soft passionate Movement, always set in a slow Triple . . . 
apt to move the Passions and to disturb the tranquillity of the Mind/ 

(504) Masson, loc. cit. 9 1701 : 

*The Saraband fis taken] gravely (gravement).' 

(505) Quantz, loc. czY., 1752: 

'The Sarabande has the same tempo [as the Entree, the Loure and the 
Courante], but is played with a rather more flattering expression.' 


The name derives from Tambour or tabor (drum), since in its Pro- 
vengal origins the dance was accompanied by a pipe and tabor. Its 
characteristic feature is a certain remorseless ferocity of rhythm, in 
which the beats of a drum partake when the music is orchestrated, and 
must be counterfeited when no drum is present. The pulse is a quick 
two in a bar. 

(506) Quantz, loc. cit., 1752: 

'A Tambourin is played like a Bourree or Rigaudon, only a little 


A late sixteenth-century dance particularly fashionable at the court 
of Elizabeth in England; she was portrayed dancing it with Robert 
Dudley, Earl of Leicester, who died in 1588. The painting is at Penshurst 


Place, Kent. It shows the typical of the by the 

woman leaps with a turning motion, by the on 
her waist, at each climax of the tune. To do 

perfect timing from the dancers, and a very exact from the 

musicians. This tempo is, in effect, a two-pulse 6-4 of a very 

steady speed, but with a crotchet unit, and therefore the 

gulliard, which has a minim unit (3-2). But Is the impression 
of an unexpectedly slow pulse supporting a very energetic 
(507) Morley, he. a/., 1597: 

'Like unto [the bransle double] (but more light) be the voltes and 
courantes which being both of a measure, are notwithstanding daunced 
after sundrie fashions, the volte rising and leaping ..." 


Quantz' speeds for dance tempos are less hypothetical and more 
actual than his speeds for time-words (see Ch. XXXVI, 4 above); but 
the variable tempos of dances at different times and places, or when 
they have become musical forms in their own right, are a factor which 
no such table can take into account. The equivalent metronome 
markings are as follows. 

TABLE VIII: Quantz" dance tempos 

Bourree [(for 2]: J = 160 

Canarie: J. = 160 

Chaconne: J = 160 

Courante: J = 80 

Entree: J = 80 

Furie: J = 160 

Gavotte: about J = 120 

Gigue: J. = 160 

Loure: J = 80 

Marche[<Por2]: J = 160 

Menuet: J = 160 

Musette [3-4]: J = 80 

[3-8]: /= 80 

Passecaille: about J = 180 

Passepied [3-4]: about J = 180 

[3-8]: about J = 180 

Rigaudon [(? or 2] : J = 160 

Rondeau [$ or 3-4]: about J = 140 

Sarabande: J = 80 

Tambourin: about J = 180 



The above table looks more useful than it really is. Wilfrid Mellers 
(Couperin, London, 1950, App. D) prints several similar tables, also 
based on eighteenth-century sources, but giving quite different results. 
Ralph Kirkpatrick has much the best discussion, in his article on 
'Eighteenth-Century Metronomic Indications 5 (see my BibL). But he 
is not able to reach any definitive conclusions. The metronome is a 
devil's invention anyhow, being accurate in every respect except what 
matters, namely the musical circumstances of the case. 

John Guthrie, from his experienced dancer's angle, suggests for the 
Minuet of about 1650-1710 a crotchet pulse of MM. 160; thereafter, 
about double that speed, though the change was presumably quite 
gradual (Historical Dances for the Theatre, Worthing, 1949, p. 27). 


The following is worth noting, especially for its description of 
Galliard speed as Very slow\ This confirms my explanation in Sect. 1 1 
above. The steps are very lively indeed; the effect is brisk and in strong 
contrast to the measured dignity of the Pavan; but the tempo remains 
moderate. A steady beat with a powerful impetus: that is what the 
dancers require of the musicians. Even when no dance is necessarily 
involved, as in some of the more musically elaborate examples, the 
same character and the same steadiness remain desirable. 

Nevertheless, the speed of the Galliard may be much quicker than 
it looks in the notation. This is due to the normal practice of notating 
triple rhythms in notes apparently but not actually slower than duple 
rhythms. See Ch. XXXVIII, Sect. 5 and 8 below. 

(507a) Giovanni Lorenzo Oregon, Arie in stil francese a una e due 
voci, Lucca, 1698, Preface: 

"For whoever does not have some knowledge of dances, we mention 
that Gagliards are usually sung in tempo assai largo. The Minuet allegro e 
vivace. The Bore presto and alia Breve* 




(a) Our time-signatures originated in renaissance notation as a means 
of indicating the duration allotted to each note- value in proportion to 
the others. 

This was necessary because the proportions were variable. In modern 
notation, we assume a standard proportion of one to two between each 
successive note- value: one semibreve to two minims; one minim to two 
crotchets; etc. But in renaissance notation, one long might contain two 
or three breves; one breve might contain two or three semibreves; one 
semibreve might contain two or three minims. Any smaller values, 
however, were in duple proportion only, 

The relation of long to breve became known as 'mood' (modus a 
word also used for the 'time-signature' itself); the relation of breve to 
semibreve became known as 'time* (tempus this word is also used for 
a unit of time); the relation of semibreve to minim became known as 
'prolation' (prolatio). 

A triple relationship between any two of these is known as 'perfect', 
and a duple relationship is known as 'imperfect*. 

The breve-unit measure of 'time' may also be referred to as the 
'more' and the semibreve unit measure of 'prolation' as the 'less*. 
During the renaissance period, 'mood' became obsolete in practice, 
because of the perpetual tendency for long notes to grow slower as 
shorter notes are added at the other end of the scale. 

Our name for this system of mensuration as a whole is proportional 
notation. For English readers, the best and most accessible contem- 
porary account of it, written just as it was becoming obsolescent, is in 
Morley's Plaine andEasie Introduction (London, 1597) in the annotated 
modern edition mentioned in my Bibliography. The best short account 
by a modern English authority will be found in two articles by Thurston 
Dart (Grove's Dictionary of Music and Musicians, London, 1954, VoL 
VI, s.v. 'Notation' and 'Proportions'). 

(b) At no period were the signs and symbols of proportional notation 
brought to a consistent or systematic condition; but a few of the best 
entrenched around 1600 were as follows. 



TABLE IX: of Mensuration c. 160G 


O Perfect: \&\ = & a e> Perfect: = d J J 

O Perfect: M = o o ^ Imperfect: G = J J 

G Imperfect: \o\ *=* o & Perfect: = d J J 

C Imperfect: |o| = o o Imperfect: & = J J 


(]) Note values become half those of O 

CorO c 

Cor 5> C 

| Dupla (diminution): J = previous o 

][ Dupla (augmentation): & = previous J 

I or ^ Tripla (diminution): 000 = previous c? 

| or 3 Sesquialtera (diminution): J d J = previous o 

(c) For ail duple relationships, our modern system, with its one 
standard proportion, is on grounds of simplicity an obvious improve- 

For triple relationships, the improvement is by no means so obvious. 
In order to provide a note containing three notes of the next shortest 
value, we have to fall back on one of the complications of the old 
proportional system: the dot of augmentation. We have to say: let a 
dotted crotchet, etc., contain three quavers, etc. 

But for combining duple with triple relationships, we need two kinds 
of quavers, etc.: our standard kind, of which two go to a crotchet; and 
a further kind, of which three go to a crotchet. We have to say: let three 
quavers, etc., written where there is apparently only time for two, be 
this further kind of quaver, of which three go to a crotchet. We call this 
triplets. But when we write triplets in duple time, or duplets or quad- 
ruplets in triple time, or quintuplets or sextolets or septuplets, etc., we 
are no longer using our modern uni-proportional notation at all. We 
are using a remnant of the old multi-proportional notation, and using 
it very sensibly. 

To indicate the proportion, we commonly write 3 against the triplet, 
etc., often adding a square bracket to make the grouping clear if there 
are no connecting beams to make it clear already. But the proportion 
required may be so obvious that even this indication may not be 

(d) In renaissance music, however, the proportion might not be 
obvious at all, especially where different parts in the same polyphonic 
composition were using different proportions, or where the proportion 



In course of a composition. It was to use 

symbols or numerical signs of mensuration. 

These are the symbols and signs we have the 

name of time-signatures. 

From what has been said, it will be this is some- 

what misleading. The true function of these signatures was not to 
indicate time in our modern sense of tempo, but to indicate time in the 
old sense of mensuration. The only information which they Impart 
directly is the relative time allotted to each note- value in proportion to 
the others. Any information which they may incidentally impart about 
time in the absolute is indirect information, and neither exact nor 

The only reason why the signatures can impart even indirect informa- 
tion about the absolute time in which the music proceeds is that for 
certain stable forms of late renaissance polyphony a conception seems 
to have been current of a roughly uniform or standard pulse (tactus) 
of which the remaining note-values were either multiples or sub- 
divisions. It will be appreciated that this was not a means of dictating 
tempo, which would be a musical absurdity, but of teaching it, and that 
it would not have been practicable if it had not been kept quite flexible 
in its actual application. 

(508) Thomas Morley, Plains and Easie Introduction, London, 1597, 
p. 9: 

*[A stroke] is a successive motion of the hand, directing the quantitie of every 
note and rest in the song, with equall measure, according to the varietie ofsignes 
and proportions. . . . The More stroke they call, when the stroke comprehendeth 
the time of a Briefe. The lesse, when the time of a Semibriefe: and proportional 
where it comprehendeth three Semibriefes, as in triple or three Minoms, as in 
the more prolation. . . . 

'[The timing of a note] is a certayne space or length, wherein a note may be 
holden in singing. 

[p. 23:] 'The signe of this Moode set with a stroke parting it thus (C causeth 
the song before, which it is set, to be so song as a breefe or the value of a breefe 
in other notes, make but one fid stroke, and is proper to motetes specially when 
the song is prickt in great notes. . . . Although that rule bee not so generally 
kept: but that the composers set the same signe before songs of the semi- 
brief e time.' 

It would, of course, be for the hand to accommodate itself to the 
natural tempo of the music, not the music to the hand. But when the 
notation is, throughout all the parts, in note-values which look longer 
than they are meant to be performed, Morley tells us that a < will 
warn the singers to read them as if they were notes of half the written 
value: i.e. to treat each of their parts as they would if some of those 



parts were notated in notes of half the written value under the signature 
C. The result is, at least in theory, that they all sing faster than they 
would if all their parts were written in the same values but under the 
signature C, This is the same result as we might secure by adding 
'allegro* or 4 non troppo lento'; but it has been reached by an entirely 
different process of thought. 

In practice, as opposed to theory, the singers would have produced 
the same result by trained musicianship and good judgment of tempo, 
whether they saw C or C or neither. That is why, as Morley at once 
adds, composers did not always trouble to distinguish C from $ in such 
cases: Le. where all the parts are in the same proportions, and remain 
so. The matter is quite otherwise where C appears in some parts but (P 
in other parts, or where there is a change in course of the composition. 
There the signatures are necessary in order to keep the relative time- 
values in correct proportion; but then there is not even an indirect 
influence on tempo in the absolute. 



(a) During the early baroque period the proportional system de- 
clined. Its place was taken by our present uni-proportional system, of 
which the normal basis is duple. For this reason, it no longer requires 
the signs of imperfection (duple relationship) such as C and <p: for this 
is assumed except where perfection (triple relationship) is shown or 

(509) Pierre Maillart, Les Tons, Toumai, 1610, p. 349: 

'When there is no sign of perfection, the usage has established itself 
that to signify a general imperfection, a half circle is ordinarily used, as a 
mark or sign of imperfection: in that, just as the round figure, or the 
perfect circle, is the true hieroglyphic, and the essential mark of perfection, 
so the imperfect circle, is the mark of imperfection. ... It follows from 
what we have said, that the signs of imperfection dealt with above, are 
superfluous and useless, seeing that the absence of signs of perfection 
suffices to declare the imperfection, as has been said/ 

(b) Unfortunately the superfluous signs remained in use, but became 
subject to a fundamental confusion. 

This is the confusion of relative time-values with tempo in the 

It was remembered, for example, that the stroke of $ stands for 
diminution of C by one half. 

This had meant: take two notes in the time-value of one, relative to 
a previous C or to a C continuing in another voice of the same 



It was now explained by the more as 

take an entire composition at twice the The less pedantic, 

knowing that this did not happen in practice, compromised with some 
such evasion as: take a somewhat faster absolute speed. Only the really 
learned, however, could put the matter as realistically as the following. 

(510) Athanasius Kircher, Musurgia, Rome, 1650, p. 676: 

'There is nothing in music more confused ... [p. 679] this most con- 
fused subject ... this utter muddle (tota haec farrago), [p. 679:] an 
upright line [through the sign means that] the notes must be halved, that 
is to say, all the voices must sing twice as fast in those parts which are so 
marked, [p. 684: but we could nowadays show this] by the rapidity of our 
notes, such as minims, crotchets, quavers, semiquavers; hence we too 
judge it superfluous to use those signs; indeed I have found that a majority 
of the most excellent musicians and the most expert in theory of the 
present time have deliberately omitted them, and taken them for one 
and the same sign (pro unico signo).' 

Zarlino (Institution! Harmoniche, Venice, 1558, ed. of 1562, p. 278) 
had already made an almost identical attack. Where such great author- 
ities agreed, who are we to expect reliable tempo indications from C, 
(C, etc.! See also quotations (518) to (519) below. 


A figure standing as a proportional time-signature shows the number 
of units to be taken in the time of one previous unit. Where there are 
two figures, the first or the top figure (numerator) shows the number of 
units to be taken in the time of the number shown by the second or the 
bottom figure (denominator). 

Thus 3 or ^ or i means take three units in the time of one: i.e. in 
the proportion of 3: 1. \ means take three units in the time of two: Le. 
in the proportion of 3:2; etc. This, in its context, was self-explanatory. 

As the context became forgotten, so did the explanation. It came to 
be assumed that the top figure gives the number of units to a bar, and 
the bottom figure the size of these units in relation to the semibreve. 

Thus now means four quarters of a semibreve, i.e. crotchets, to a 
bar; \ two crotchets to a bar; \ four quavers; \ three crotchets; | three 
quavers, etc. 

By the time this changed theory was recognised, perhaps first in 
G. M. Bononcini, Musico prattico, Bologna (1673), the facts themselves, 
by one of the pleasanter ironies of history, had changed to fit it; and 
so convenient is it that it has always since been taught. 

(511) Leopold Mozart, Violinschule, Augsburg, 1756, 1, ii, 5: 

"Among the time-measures, common time is the chief measure, into 



relationship with which all the remainder are brought: for the top number 
is the numerator while the bottom number is the denominator. We may 
therefore say that of those notes of which four go to a bar of common 
rime, two go to a bar of J time. From this we see that \ time has only two 
parts, the up beat and the down beat, and since four black notes or 
crotchets go to common time, therefore two of the same value go to \ 
time. In this manner all time measures are calculated.' 


(512) John Playford, Introduction to the Skill of Musick, London, 

1654, p. 15: 

'{Recent composers] make use onely of . . . the Triple Time [and] the 
duple, or Common Time . . . [p. 17:] thus marked <t, and is usuall in 
Songs, Fantasies, Pavins, Almans, and the like. [Not pre-1672:] Note, 
That when this Common Mood is reversed thus 3), it is to signifie, that the 
Time of that Lesson or Song, before which it is so set, is to he Playd or Sung 
as swift again as the usual Measure. 

'[1694 on, "Corrected and Amended by Mr. Henry Purcell", p. 25:] 
First, I shall speak of Common-Time, which may be reckond three 
several sorts; the first and slowest of all is marked thus C: J Tis measured 
by a Semibreve, which you must divide into four equal Parts, telling 
one, two, three, /our, distinctly, putting your Hand or Foot down when 
you tell one, and taking it up when you tell three, so that you are as long 
down as up. Stand by a large Chamber-Clock, and beat your Hand or 
Foot (as I have before observed) to the slow Motions of the Pendulum . . 

'The second sort of Common Time is a little faster, which is known by 
the Mood, having a stroak drawn through it, thus (p. 

'The third sort of Common Time is quickest of all, and then the Mood 
is retorted thus 3) ; you may tell one, two, three, four in a Bar, almost as 
fast as the regular Motions of a Watch. The French Mark for this retorted 
Time, is a large Figure of 2! 

(513) Henry Purcell (or his editor), Lessons, 1696, preface: 

'There being nothing more difficult in Musick then playing of true 
time, 'tis therefore necessary to be observed by all practitioners, of which 
there are two sorts, Common time and Triple time, Si is distinguished by 
this C this < or this 3) mark, ye first is a very slow movement, ye next a 
little faster, and ye last a brisk & airry time, & each of them has allways to 
ye length of one Semibrief in a barr, which is to be held in playing as long 
as you can moderately tell four, by saying one, two, three, four/ 

(514) Anon, Compleat Flute Master, London [c. 1700], cited A. 
Dolmetsch, Interpretation, London [1915], p. 34: 

*C very slow motion, (p somewhat faster. 3) Brisk and light Ayres/ 



(515) Dean, Complete Tutor for the Violin* London, 1707, A, 
Dolmetsch, Interpretation* London [1915], p. 34: 

*C Very solid or slow movement. C Quicker, b or 2 as quick as 

the first, and arc calFd Retorted Time/ 

(516) Christopher Simpson, Compendium of Music (London, 1665, 
as Principles of Practical Mustek)* p. 19: 

'The Signe of [the Common] Mood is a Semicircle, thus, C, sometimes 
with a Stroke through it thus C. 

[Ed. of 1706, p. 12:] 'The Sign of [the Common] Mood is a Setnicircfa 
thus, C, which denotes the slowest Time, and is generally set before grave 
Songs or Lessons; the next is this C which is a Degree faster, the next 
mark thus 3) or, thus 2, and is very Fast, and denotes the Quickest 
Movement in this Measure of Common Time. [Likewise to 9th ed. of 

[p. 22:] 'The motion of the Hand is Down, and Up, successively and 
equally divided. Every Down and Up, being called a Time, or Measure. 
And by this we measure the length of a Semibreve [and from it compute 
the other note-values] . . . 

'But you may say: I have told you that a Semibreve is the length of a 
Time, and a Time the length of a Semihreve, and still you are ignorant 
what that Length is. 

'To which I answer, (in case you have none to guide your Hand at the 
first measuring of Notes) I would have you pronounce these words (One, 
Two, Three, Four) in an equal length, as you would (leisurely) read diem. 
Then fancy those four words to be four Crotchets, which make up the 
quantity or length of a Semibreve, and consequently of a Time or Measure: 
In which, let those two words (One, Two) be pronounced with the Hand 
Down; and (Three, Four,) with it Up. In the continuation of this motion 
you will be able to measure, and compute all your other Notes. 

'Some speak of having recourse to the motion of a lively pulse for the 
measure of Crotchets; or to the little Minutes of a steddy going Watch for 
Quavers; but this which I have delivered, will (I think) be most useful to 

Like Quantz' and other early instructions for turning relative time- 
values into absolute tempos, the above was not meant to limit the 
beginner's choice of tempo, but to give him an average starting-point 
6 in case you have none to guide your hand at the first measuring of 
notes*. Arnold Dolmetsch (Interpretation., London [1915], p. 31), 
taking the mean pulse rate at 75 since that 'makes the quavers agree 
with the "little minutes" or strokes of a watch, which usually beat five 
times a second or 300 times a minute', makes this MM J = 75 : reason- 
able enough as an ordinary allegro, but it is only a starting-point. 



(517) Georg Muffat, Florilegium /, Augsburg, 1695, preface (here 
translated mainly from his parallel text in French): 

'The measure marked thus 2 [or] C, being given in two beats, it is clear 
that in general it goes as fast again, as this C which is given in four. It is 
however understood that this measure 2 ought to go very slow in Over- 
tures, Preludes, and Symphonies, a Ettle more lively in Ballets, and for 
the rest on my advice almost always more moderate than this <, which 
itself ought to be less pressed on in Gavottes, than in Bourrees. However 
when this measure 2 is given very slowly, and (as has been said) in two 
beats, the notes are almost of the same value, as with the Italians under 
this measure C given in four beats with speed under the word presto' 

(518) Charles Masson, Nouveau Tralte, Paris, 2nd ed., 1701, p. 6: 

'Measure is the soul of music, since it excites with such truthfulness of 
emotion (fait agir avec tant de justesse) a great many people, and by the 
variety of its times (rnouvemens) can again stimulate so many different 
feelings, being able to calm these and arise those, as has always been 

'Although there appear to be a quantity of different measures, I believe 
that it is useful to point out that it is only the number two or three which 
divides them, and that it is by the quickness or by the slowness of these 
two times (mouvemens) that the difference between melodies is effected/ 

(518a) fitienne Loulie, Elements . . . de musique, Paris, 1696, p. 69: 

'All these time-signatures were in use among the ancients . . . foreigners 
have retained some in their works, but the practice of them is not very 
certain, some use them in one way, some in another.' 

(519) Michel de Mont<clair, Petite Methode, Paris, [c. 1730], p. 48: 

'There is a time-signature marked by (p, of which the usage is no 
longer well defined; it is used in different manners for lack of willingness 
to recognise its character/ 

(520) Le Cerf de La VieviUe, 'Comparaison', in Histoire de la 
Musique, Paris, 1725, I, p. 307 (also reprinted and translated in 
Mattheson's Critica Musica, III, pp. 222 ff.): 

'[The Italians] find no music good, if it is not difficult, they can scarcely 
bring themselves to look at it when there are only minims, or crotchets in 
two- or three-time; as if all the Italian measures did not come down to 
these two measures, as if one could not reduce them to two- or four-time, 
and include two measures in one, as if the 4-8 did not come down to our 
light two-time? and the 6-8 measures, the 3-8 and the 12-8 did not come 
down to our three-time measure? when they are beaten more or less 
quickly, whether they are beaten in two-time or in four-time, of which 
each beat makes up one of our measures in three-time? This is nothing but 



a different manner of expressing what is in itself, and the 

character of the piece, for slowness quickness, has con- 

venience for beating; for as there are in general only different modes, 
the minor and the major, so there are in general only two 
in two-rime, and that in three ; in vain would you wish to imagine others.* 

(521) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. 

IV, Sect. 48, p. 348: 

'The Overture Time is generally indicated by 2 2 . . . 

[49:] 'Its measure is properly slow and expressive . . . 

[50:] 'But if in certain pieces the Overture Time is meant to have a 
rapid movement, then, instead of the slow measure just mentioned, a 
stroked 2 or ( is shown, or rather, with a few people it would be shown, 
in distinction from the slower measure. But since such correctness is not 
always found, and since the signs 2, 2, C or C are used without dis- 
crimination, sometimes for a naturally rapid piece and sometimes for a 
slow one, [(? may often serve for the slow introduction of the overture 
as well as for its succeeding quick movements]/ 

(522) Alexander Malcolm, A Treatise of Mustek, Edinburgh, 1721, 
p. 394: 

'COMMON TIME is of two Species, the 1st where every Measure is 
equal to a Seniibreve ... the 2d, where every Measure is equal to a 
Minim. . . . The Movements of this Kind of Measure are very various; 
but there are Three common Distinctions, the first is slow, signified at the 
Beginning by this Mark C, the 2d is brisk, signified by this (J5 , the 3d is 
very quick, signified by this 3) ; but what that slow, brisk, and quick is, is 
very uncertain, and, as I have said already, must be learned by Practice. 
[The usual references to a watch, or to counting One, Two, Three, Four, 
or to a pendulum, are then given for what they are worth.]' 

(523) William Turner, Philosophical Essay on Mustek, London, 1724, 
p. 20: 

*C which denotes the slowest Movement . . . 

*<C which denotes the Movement to be somewhat faster than the former. 

*3) which is the quickest of aU; the Crotchets being counted as fast as the 
regular Motions of a Watch. 

[p. 27:] 'out of all these eight Moods I have been speaking of, there is, in 
reality, but One ... the Reasons for the others, I cannot well under- 
stand . . / 

(524) Joachim Quantz, Essay, Berlin, 1752, V, 13: 

*In four-crotchet time it must be carefully observed that when a stroke 
goes through the C ... such a stroke signifies, that all the notes, so to 
speak, become of a different value, and must be played as fast again, than 



is the case when the C has no stroke through It. This measure Is called : 
allabreve, or alia capclla. But since with regard to the aforesaid measure 
many have fallen into error through ignorance : it is most desirable that 
everyone should become acquainted with this difference. This measure is 
more frequent In the galant style, than It used to be in former times.' 

(525) Leopold Mozart, Violinsckule, Augsburg, 1 756, I, ii, 3 : 

'In the old music [of the time of Glareanus and Artusi, i.e. 16th cent.] 

there were conflicting views [as to notation], and everything was in 
great confusion. 

[4:] 4 Nowadays time Is divided into equal [simple or common] and 
unequal [triple] measure [of which the equal are:] 

'The Equal Time-Measure 
C 2 or I C 

The straight or The two-crotchet The Allabreve 

four-crotchet time 


[7:] 'This Is only the ordinary mathematical division of the bars, 
however, which we call the time-measure and the beat. Now comes a 
major issue, namely the source of the speed. Not only have we to beat 
time correctly and steadily, but also to be able to divine from the music 
itself whether it needs a slow or a somewhat faster speed.' 


Triple time-signatures tend to be diminutions rather than augmenta- 

Baroque triple time-signatures therefore tend to make time-values 
shorter in relation to common time: hence to increase the speed. 

This tendency of triple-time movements or passages to move faster 
than their written note-values suggest, though not reliable, is nearer to 
being reliable than the tendency of (p to move faster than C, etc. It 
grew less reliable, however, in course of the baroque period. 

The inconsistencies in the following evidence will again be noticed. 

(526) John Playford, Introduction, London, 1654, p. 16: 

* The Imperfect of the More ... is thus signed Q 9 and this is called the 
Triple Time. This Mood is much used in Ayery Songs and Galliards, and 
is usually called Galliard or Triple time and this Triple time is in some 
Lessons, as Coranto's, Sarabands, and Jigs brought into a Measure, as 
swift again [eds. of 1662 to 1687 have 'to another swifter motion'], for 
as before three Minims or Sembriefs [sic] with a prick [dot] made a 
Time, in this three Crochets [sic] makes a Time, or one Minim with a 



prick, and tSiis measure is knowne by this signc or mark J1 [3-1 j is 

usually called Three to one. 

[1694 (p. 26) on: | "Tlicrc arc two other sorts of Time be 

reckoned amongst Common-Time for the equal division of the Bar with 
the Hand or Foot up and down: The first of which is Six ID 

each Bar containing six Crotchets, or six Quavers, three to be with 

the Hand down, and three up, and is marked thus 4, but very brisk, and is 
always used in Jigs. 

'The other sort is called Tiwlve to eight, each Bar containing twelve 
Quavers, six with the Hand down, and six up, and marked thus ! 8 2 . . . 

4 Tripla- Time, that you may understand it right, I will distinguish into 
two sorts: The first and slowest of which is measured by three Minims in 
each Bar, or such a quantity of lesser Notes as amount to the value of three 
Minims, or one Pointed [dotted] Semibreve, telling one, two, with your 
Hand down, and up with it at the third; so that you are as long again 
with your Hand or Foot down as up. This sort of Time is marked thus i 

'The second sort is faster, and die Minims become Crotchets, so that a 
Bar contains three Crotchets, or one Pointed Minim; 'tis marked thus 3 
or thus ^1 [3-1]. Sometimes you will meet with three Quavers in a Bar, 
which is marked as the Crotchets, only Sung as fast again. 

'There is another sort of Time which is used in Instrumental Mustek, 
called Nine to six, marked thus I, each Bar containing Nine Quavers or 
Crotchets, six to be Play'd with the Foot down, and three up: This I also 
reckon amongst Tripla-Time, because there is as many more down as up.' 

These explanations are given as proportions. By the next (thirteenth) 
edition of 1697, the signature 9 6 (as 9:6) has been replaced by | (as nine 
quavers) in tacit acknowledgement of the later reckoning by fractions 
of the semibreve. 

(527) Christopher Simpson, Compendium, London, 1665, I, 10, Of 
Tripla Time : 

'When you see this Figure [3] set at the beginning of a Song or Lesson, 
it signifies that the Time or Measure must be compted by Threes, as we 
formerly did it by Fours. 

'Sometimes the Tripla consists of three Semibreves to a Measure, each 
Semibreve being shorter than a Minim in Common Time . . . 

'The more Common Tripla is three Minims to a Measure, each Minim 
about the length of a Crotchet in Common Time . . . 

'In these two sorts of Tripla, we compt or imagin these two words 
[One, Tivo] with the Hand down, and this word [ T/zree] with it up . . , 

[1706 (p. 23) on:] 'Sometimes the Tripla consists of three Minims to a 
Measure. The more common Tripla is three Crotchets to a Measure . . , 

[p. 37:] 'There are divers Tripla s of a shorter Measure; which by 
reason of their quick motion, are usually measured by compting three 
down and three up with the Hand. [1706 (p. 25) on:] and those quick 



Triplas arc prick* t sometimes with Crotchets and Mitwtnsi and sometimes 
with Quavers and Crotchets.* 

(528) Jean Rousseau, Methode Claire ', Paris, 1678, ed. of Amsterdam, 
1691, p. 31: 

'With the ternary signature (au Signe Trinaire) [C3, the bar is beaten] 
in three slow beats, two down, and the other up. 

'With the plain triple signature (au Signe de Triple simple) [3] in three 
quick beats, two down, and the other up. 

'With the double-triple signature (au Signe de Triple double) [|], in 
three slow beats, two down and one up; [also called] Three to Two . . . 

* With the signature of three to four, so called because in place of the 
bar composed of four crotchets, this has only three, the bar is beaten in 
three beats quicker than simple triple . . . 

'With the signature of three to eight composed of three quavers, 
where the major has eight, the bar is beaten like that of three to four, but 
very much faster. 

'With the signature of six to four [six crotchets] . . . the bar may be 
beaten in two . . . 

'With the signature of six to eight [six quavers] . . . like the six to four 
but quicker, or like the three to eight, making two bars of one. 

'With the signature of four to eight [four quavers] the bar is beaten in a 
very quick two.' 

Grassineau is interesting for his retention of C and < in front of the 
numerical signatures: e.g. C | giving three beats to a dotted semibreve, 
but <^ giving three beats to a dotted breve. His most important 
contribution, however, is the following, with which we may compare 
Simpson, second and third paragraphs at (527) above. 

(529) J. Grassineau, Musical Dictionary, London, 1740, 'Allegro': 

*It is to be observed, the movements of the same name as Adagio, or 
Allegro, are swifter in triple than in common Time/ 

The following, very possibly by Purcell, is disappointingly incomplete 
even for his own music, since his sonatas include Largos in \ for which 
the Very slow' tempo stated below is musically too slow, as well as 
conflicting with his own statement given at (435), Ch. XXXVI, 3 above 
that Largo is 'a middle movement' (i.e. neither fast nor slow); while 
some of his Hornpipes, marked |, are certainly fast. (See (533) below 
for a vivace |.) 

(530) Henry Purcell (or his editor), Lessons, London, 1696, preface: 

'Triple time consists of either three or six Crotchets in a barr, and is to be 
known by this |, this ^1 [3-1], this 3 or this niarke, to the first there is 
three Minums in a barr, and is commonly pky'd very slow, the second 
has three Crotchets in a barr, and they are to be play'd slow, the third has 



ye same as ye former but is play'd faster, yc last has six Crotchets in a 
& is Commonly to brisk tunes as Jiggs Paspys 

(531) Anon, London [c. A, 
Doimetsch, Interpretation^ London [1915], p. 34: 

s ? Grave movement. 3 Slow. J Fast, for Jiggs ? Pasples, &/ 

(532) Dean, Complete Tutor for the Violin, London, 1707, cited 
A. Dolmetsch, Interpretation, London [1915], p. 34: 

*$ Very slow. \ \ Much Quicker/ 

(533) A. Malcolm, Treatise of Mustek, Edinburgh, 1721, p. 394: 

'Movements of the same Name, as adagio or allegro, etc. arc swifter in 
triple than in common rime [the same as Grassineau at (529) above] . . 
in the triple, there are some species that are more ordinarily [N.B. not 
exclusively] of one Kind of Movement than another: Thus the triple \ is 
ordinarily adagio, sometimes vivace; the 4 is of any Kind from adagio to 
allegro; the | is allegro, or vivace; the 4 f f are more frequently allegro; the 
8 2 is sometimes adagio but oftener allegro. Yet after all, the allegro of one 
Species of triple is a quicker Movement than that of another, so very 
uncertain these Things are/ 

(534) W. Turner, Philosophical Essay on Mustek, London, 1724, p. 22: 

'[2] differs from [4] no otherwise than in being measured by different 
Notes: For in the former Mood, Minims are sometimes, as fast as Crotchets; 
and in this, the Crotchets are often as slow as Minims [while f] is exactly 
the same with the others: For in this Mood, the Quavers are sometimes, as 
slow as Minims are in the first/ 

Leopold Mozart's list is very modern, as was Ms list of common 
times. In a footnote, he adds that he would have spared us f if it did 
not occasionally appear in an old church piece. He calls J, f , g, ? 6 , }g, 
\l, ^ all 'worthless stuff' because everything necessary can be expressed 
without them. He advocates *| in preference to | for rapid melodies 
solely because it is easier to beat. 

(535) Leopold Mozart, Violinschule, Augsburg, 1756, 1, ii, 4: 

'The Unequal Time Measure 

The Semibreve The minim The three- The three- 

Triple time triple time crotchet time quaver time 

6 6 12 

The six- The six- The twelve- 

crotchet time quaver time quaver time* 




Discrepancies often occur in the time-signatures given by different 
contemporary copies of the same music, or even between different parts 
in the same copy, etc. See Walter Emery (Music and Letters, XXXIV, 
iii, July, 1953, review) for a revealing study of such discrepancies in 
J. S. Bach. The inference, as usual, is not that tempo was disregarded, 
but that the performer must find it in any case; while the time-signature, 
being at best of little help, received only casual attention. 


In addition to inadvertent discrepancies, we find a few genuine 
survivals of proportional notation. 

A melody may be in | against a bass in *, or a melody in ! 8 2 against 
a bass in C, i.e. \, etc. One crotchet of the bass then goes to three quavers 
of the melody, as if that melody were notated in triplets. If the bass 
presently requires shorter notes, these may actually be notated as 
(unmarked) triplets, or alternatively with dots intended to be softened 
into triplet rhythm. In spite of various illogicalities here, no real 
obscurity usually arises if the appropriate conventions are known, for 
which see Ch. XLVI below. 

L S. Bach's Chorale Prelude 'Herr Gott, nun schleuss' as printed in 
the Bach-Gesellschaft (25, ii, p. 26) has C on the top stave, C on the 
middle stave and C l j- on the bottom stave. This looks formidable; 
but it merely states that the middle line has sextolets while the bottom 
line has triplets against the melody in the top line. If it were printed 
C in all three parts, or nothing at all, this would still have been obvious. 
It could then, as a precaution, have printed 6 against the first few 
sextolets and 3 against the first few triplets, though this precaution 
would not actually have been needed. 

Walter Emery pointed out to me, however, that Bach himself wrote 
this music, like most of his organ music, not on three staves but on 
two. The top has C ^ and the bottom has C 1 8 2 . The upper melody 
proceeds, in minims, crotchets, etc., exactly as if its time-signature had 
been C, which in fact it is not. The tenor counter-melody, in sextolet 
relationship to the upper melody, wanders from stave to stave with a 
sublime disregard for their having different time-signatures. The bass, 
in triplet relationship to the upper melody, goes on unperturbed; but 
at bar 19 the alto part is written as duplet quavers. From their harmony 
these can only be played unevenly, in a triplet rhythm to fit the bass 
(see Ch. XLVI below). 

In the third verse of J. S. Bach's Chorale Prelude 'O Lamm Gottes 
unschuldig' (B.G. 25, ii, p. 106) the top stave is changed from \ to \ 



while the bottom stave remains in \. This the of the top 

two parts into triplet relationship with the of the 

again the part wanders regardless between the two staves. 

A more tortuous case occurs In Prelude XV from J, S. Bach's 4 Forty- 
Eight*. Here we start with what in effect are semiquaver triplets in the 
treble accompanied by quavers in the bass. This J. S. announces 

in the grand manner by setting J in the treble stave and C in the 
stave. Presently he wants, very reasonably, to reverse his counterpoint 
by putting the triplet semiquavers in the bass and the quavers In the 
treble. He does so; the musical effect is excellent. But what about the 
effect on the notation? 

There is no effect on the notation. The time-signatures remain un- 
changed. The semi-quavers in the bass are now unmarked triplets in 
theory as well as in practice; the quavers in the treble (still, remember, 
in \l) are proportional quavers which it would be quite complicated to 
work out mathematically, but for the fact that It is totally unnecessary 
to do so. The performer knows at once that they are the same quavers 
as he has had previously in the bass. We could not have a more striking 
instance of the unimportance of notatlonal detail, provided the sense 
is clear to an ordinary musical intelligence. And ordinary musical 
intelligence was precisely the quality most taken for granted in baroque 


To sum up: 

(a) Different time-signatures used concurrently can give exact infor- 
mation, but only on the relative time-values in the concurrent parts. 

(b) A sign or a number crossed or retorted suggests a faster absolute 
tempo; but this suggestion is vague and unreliable. 

Smaller denominators suggest faster tempos; but this suggestion is 
equally vague and unreliable. 

Any triple-time signature suggests a faster tempo; and this sugges- 
tion, though still as vague, is not quite so unreliable as the foregoing. 
It generally works. Often a very much faster tempo is required. 

(c) A change of signature in course of the composition is more likely 
to be of significance than the choice of signature at the beginning. 

(d) Besides tempo, time-signatures have a bearing on pulse, which 
will be considered in the next chapter. 





(a) Pulse is not the same as accent, though the two may often coincide. 

In renaissance polyphony, the accentuation follows only the natural 
shape of the phrase, not the underlying pulse. The accents in the 
different parts seldom come together, and there is no such thing as a 
regular accented beat. 'Follow the rhythm of the words, not the barring' 
is usually good practical advice. Yet the pulse, though not made audible, 
is somehow present at the back of one's mind as the groundwork 
against which the irregular accentuation takes its meaning. This is very 
like what happens so frequently in poetry: the regular stresses of the 
metre, once established, are present in one's mind; the actual stresses 
of the words partly conflict with them; and in this tacit 'counterpoint' 
lies much of the beauty. 

In baroque music, the accentuation is likely to coincide with the 
pulse much more frequently; yet there are a great many passages in 
which this appears to be the case, but is not. The accentuation still goes 
by the phrase and not merely by the bar; and to find the actual peak- 
note to which the phrase is leading is often most difficult and subtle 
where the notes present a regular appearance. 

(536) Francesco Geminiani, Art of Playing on the Violin, London, 
1751, p. 9: 

'If by your manner of bowing you lay a particular Stress on the Note 
at the Beginning of every Bar, so as to render it predominant over the 
rest, you alter and spoil die true Air of the Piece, and except where the 
Composer intended it, and where it is always marked, there are very few 
Instances in wlrich it is not very disagreeable/ 


There are no sharp distinctions; but if the harmony changes often 
enough for us most naturally to count four in a unit, we feel a four- 
square pulse; if the harmony changes infrequently enough for us most 
naturally to count two in a unit we feel a swinging pulse like a pendulum. 

The unit may be a bar, or two bars thrown into one, or half a bar: 
this affects only the notation, not the music. Similarly we may for 
convenience of conducting beat a fast four-pulse unit in two, etc. 



(537) Michel Corrette, de la Paris-Lyons, 
c. 1730 9 p. 4: 

*The measure of 4 beats ... is in two different ways; to wit, in 

the Allegro always once [up and down] : in the Adagio, or other slow 
pieces, twice if you so desire.* 

C or 2 in place of C, etc., could be used to indicate a two-pulse unit 
In place of a four-pulse unit; so could the terms alia breve, alia capellaj 
mesure a deux temps, etc. ; but the fatal confusion into which the system 
fell in practice can be seen below. 

(538) Jean Rousseau, Methode Claire . . . pour & chanter ^ 
Paris, 1678, ed. of Amsterdam, 1691, p. 30: 

'With the major signature [C] the bar is beaten in four steady beats, 
two down and two up. 

'With the minor signature [C], it is beaten in two slow beats, one down, 
and the other up. 

'With the binary signature [2], in two quick beats, one down, and the 
other up.' 

(539) Sebastien de Brossard, Dlctionaire de Muslque 9 Paris, 1703, 
s.v. 'Tempo': 

'The plain C appears in two ways; 1. turned from left to right, thus 
C, and then die Italians call it Tempo ordinario, because it is in more 
ordinary use than any other; or again Tempo alia Semibreve, because 
under this signature a Semibreve or Ronde o is worth one bar or four 
beats and the other notes in proportion. 2. But it is sometimes found 
turned from right to left thus 0, when all the figures are diminished by 
half their value. Thus a Ron de [Semibreve] is only worth two beats, a 
Minim or Blanche is only worth one beat, and so with the others. 

'The stroked C is found, also either turned from left to right thus (?, 
or from right to left thus $. When it is turned to the right the Italians, 
again, call it Tempo alia breve, because in earlier times all the notes were 
diminished under this signature by half their value; but now it indicates 
that the bar should be beaten in two steady beats, or in four extremely 
quick beats; at least if it has not got Largo, Adagio, Lento or some other 
term which tells us that we must beat the bar extremely slowly. And 
when the words Da Capella, and alia breve, are seen with this signature, 
this indicates two very quick beats. 

'Finally others again of the more modern [musicians] divide time into 
only two varieties, The first is Tempo Maggiore, or Major Time which is 
indicated by a $ stroked and signifies that all the Notes can be sung alia 
breve, that is to say by making them last only half their ordinary value. 
The second is Tempo minore or Minor Time which is indicated by a plain 
C and under which all the notes have their natural value.' 



J. G. Walther Lexicon, Leipzig, 1732, s.v. 'Tempo') and 

Qoantz (Essay, Berlin, 1752, XVII, vii, 50) give the same names, 
signatures and descriptions as the last paragraph above; it will be 
noticed that Brossard at (539) above reverses the names given by Jean 
Rousseau at (538), but his signs and descriptions are the same. Leopold 
Mozart, on the other hand, at (540) below, gives the same names as 

(540) Leopold Mozart, Violinschule, Augsburg, 1756, 1, ii, 6: 

'Allabreve is a diminution of even rime. It has only two pulses, and 

is no more than J time divided into two. The signature for Allabreve is 
the letter C with a stroke drawn through it: (P. In this measure only a few 
ornaments are introduced. 
[F.n.:] *The Italians call even time tempo minore; but allabreve tempo 



(a) A further complication is introduced by the terms alia breve and 
alia capella (or da capella, or a capelld). 

The term alia breve (It.) can be translated literally, 'in shortness', or 
technically, 6 in connection with the breve'. It has two main meanings 
in baroque music: first, to indicate a style of composition; second, to 
indicate a variety of pulse. 

The term alia capella (It.) can be translated c in the church style', and 
also has two main meanings, of which the first does not concern us 
here: first, to indicate voices unaccompanied or only accompanied in 
unisons; second, as the equivalent of alia breve. 

(b) The alia breve style was regarded in the eighteenth century as 
standing at the opposite extreme from the galant style. The following 
is an excellent description. 

(541) J. D. Heinichen, General-Bass, Dresden, 1728, Part I, Ch. IV, 
Sect. 38, p. 333, fji.: 

'This antique, expressive style is surely the most beautiful and the most 
conducive, the one in which the composer can best reveal his profound 
science and correctness in composition. For the part-writing in this style 
must always be pure, its progressions and resolutions correct and far 
removed from all liberties, the melodic line sustained, with few leaps, in 
all the parts, these parts loaded with syncopations and beautiful suspen- 
sions both consonant and dissonant, and replete throughout with expres- 
sive thoughts, themes and imitations, to the exclusion of any fanciful 
characteristic. It is here that we should require a correct composer. 

'[Text, Sect. 37: The alia breve] is marked either by C or CP and differs 
from all other measures in that its own movement always remains 
unaltered, and neither Adagio nor Allegro is attached to it. 



[38:] 'Only crotchets can appear as quick notes in this 

shorter notes, namely quavers, arc allowed cither not at all or very 
sparingly, and barely two at a time. 

[45: The alia breve] divides the harmony of each bar into 2 only, 

is that not less than 2 crotchets can go to a chord/ 

The purpose of the chapter from which the above quotations are 
taken is to teach an accompanist to recognise passing notes In the bass, 
so that he does not introduce a change of harmony more often than is 
intended. By quick notes, therefore, Heinichen means notes so quick 
that they must normally be treated as passing notes when they appear 
in a figured bass. And the point of the last two paragraphs quoted is 
that in the alia breve style, these notes are crotchets, not less than two 
of which (and sometimes as many as four) should go to each change 
of harmony. 

It is perhaps worth pointing out here that this chapter from Heinichen 
has been misinterpreted by Fritz Rothschild (The Lost Tradition in 
Music, London, 1953, pp. \32ff.) in support of his theory that tempo 
in baroque music was regularly indicated and can now be discovered 
from the presence of C or (P respectively, in conjunction with the length 
of the quickest notes appearing substantially in the composition. All 
that Heinichen writes concerning which notes to take as quick for the 
purpose of detecting passing notes, Rothschild misreads as instructions 
for detecting tempo. He makes a similar misinterpretation of the 
extensive baroque evidence concerning which notes to take as quick 
for the purpose of making them expressively unequal in performance 
although written equal: for which see Ch. XLV below. He also 
believes (p. 3 and passim) that all notes on what he terms 'structural 
beats' should be audibly stressed and (p. 181) "held slightly longer than 
other notes'. Those who have heard his demonstrations will realise 
that he means both these catastrophic recommendations literally. His 
book is sincere, but dogged by misfortune, and the reader is warned to 
approach it with the greatest caution. 

We learn from Heinichen that the true alia breve, in addition to its 
'antique, expressive style', has a rhythmic characteristic connected with 
its harmony. It is characterised by a harmonic pulse of two in the bar. 

We also learn that 'its own movement always remains unaltered, and 
neither Adagio nor Allegro is attached to it'. In other words, its tempo 
is always moderate. At such a tempo, a pulse of four in the bar might 
otherwise be expected. Thus apart from its learned structure, the 
specific feature of the alia breve style is an unusually steady two in the 
bar. It has a gait like a pendulum swinging slowly. 

Its signature, according to Heinichen, may be either C or (p a very 
helpful condition of affairs! 



(c) Helnlchen next informs us of certain varieties of semi-allabreve 
which may or may not retain something of the 'antique, expressive 
style*. These condense the harmonic pulse of two alia breve bars into 
one bar of semi-allabreve, by halving the written note-values. To the 
eye, this brings back a pulse of four crotchets in the bar. But since these 
crotchets, though moving faster than the minims of alia breve, will 
certainly move much less than twice as fast, the effect to the ear may 
still be that of a slightly faster alia breve. The ear does not perceive that 
two bars have been condensed into one and the written note-values 
halved; the ear merely perceives that something of the swinging, 
pendulum-like gait remains, although not at quite so moderate a tempo. 

The signature shown by Heinichen for his semi-allabreve is C. 

(d) From the above two usages we arrive at the general use of alia 
breve, etc., to indicate not the 'antique expressive style' or any derivative 
of it, but simply a pulse of two in a bar with or without implication 
as to tempo. See Ch. XXXVIII, 4 (524) Quantz, and (525) Leopold 
Mozart; present chapter, 2, (539) Brossard, middle paragraph, and 
(540) Leopold Mozart. 


In practice none of the possible indications for a two-pulse rhythm 
can be relied on, and it is the music and not the notation from which 
we must discover it; though a change from C to C, or the appearance 
of 'alia breve' in the course of a piece, may well be a valid hint. 


When two bars of triple time unite to form, in effect, one bar of 
twice the duration, a change of pulse occurs to which the name hemiola 
(hemiole, hencdoHa, lit. one and a half) is sometimes given (the implica- 
tion is a proportion of 3 : 2). 

This change of pulse is generally though not always quite visible in 
the notation, especially of the bass part, which tends to be tied across 
the bar-line in the middle of the hemiola. It is this bar-line which is in 
effect suppressed. 

In performance, it is extremely important to make the change of 
pulse audible. Crudely: instead of one two three, one two three, the 
pulse becomes one two three one two three. 

The hemiola is mainly but not only found at (passing or final) 



Variations of Tempo 


Having decided on a basic tempo, we have to apply it with the necessary 
flexibility. Both evidence and experience confirm that the ordinary 
flexibility to which we are accustomed In later music is required for 
earlier music too, In a perfectly normal way. The next four quotations 
are quite general The practical applications of this general principle 
will be treated under separate headings below, where further evidence 
In confirmation of the principle will be found. 

(542) Thomas Mace, Mustek's Monument, London, 1676, p. 147: 

'Many Drudge, and take much Pains to Play their Lessons very Perfectly, 
(as they call It (that is, Fast) which when they can do, you will perceive 
Little Life, or Spirit in Them, meerly for want of the Knowledge of This last 
Thing, I now mention, viz. They do not labour to find out the Humour, 
Life, or Spirit of their Lessons! 

(543) Jean Rousseau, Traitf de la Viole, Paris, 1687, p. 60: 

'But genius and fine taste are gifts of nature, which cannot be learnt by 
rules, and it is with the help of these that the rules should be applied, and 
that liberties may be taken so fittingly as always to give pleasure, for to 
give pleasure means to have genius and fine taste. 

[p. 66 :] 'There are people who imagine that imparting the movement 
is to follow and keep time; but these are very different matters, for it is 
possible to keep time without entering into the movement, since time 
depends on the music, but the movement depends on genius and fine 

(544) Fr. Couperin, VArt de toucher, Paris, 1716, ed. of 1717, preface : 

'Just as there is a difference between grammar and declamation, so 
there is an infinitely greater one between musical theory and the art of 
fine playing. 

[p. 38 :] *I find we confuse time, or measure, with what is called cadence 
or movement. Measure defines the number and time-value of beats; 
cadence is properly the spirit, the soul that must be added to it.' 

(545) Joachim Quantz, Essay, Berlin, 1752, XI, 13: 

'The performance should be easy and flexible. However difficult the 
passage, it must be played without stiffiiess or constraint/ 





The prelude, like the toccata ('touch-piece"), was originally a free 
improvisation almost imperceptibly begun by the performer to check 
his tuning (on the lute, ere.), loosen his fingers, establish the tonality 
and prepare Ms own and his listeners' minds for the more formal music 
to follow. As it acquired some fluid artistic form of its own, so it came 
Increasingly within the bounds of time and measure; yet in the seven- 
teenth century, even preludes which were written down were sometimes 
written without bar-lines, and are in fact unmeasured, while those 
which have bar-lines and are measured may still require a greater 
freedom than ordinary measured music. 

(546) Fr. Couperin, UArt de toucher le Clavecin, Paris, 1716, ed. of 
1717, eight illustrative preludes: 

'{The player can choose one in the right key, in order] to loosen his 
fingers, or try the touch of an unfamiliar instrument. 

[p. 60:] * Although these preludes are written in measured time, there is 
nevertheless a conventional style to be followed. A prelude is a free com- 
position, in which the fancy follows whatever occurs to it. But since it is 
exceptional to find talents able to create on the impulse of the moment, 
those who make use of these preludes should play them in a free manner, 
without confining themselves to strict time, except where I have pur- 
posely marked it by the word measured (mesure)/ 


Another style of composition in which an exceptional freedom of 
tempo is intended is the style which started as the nuove musiche> the 
6 new music' of the Italian monodists at the turn of the sixteenth and 
seventeenth centuries. This style did not continue for more than a 
generation in its first bold and uninhibited form; but it left a sequel 
of more or less declamatory idioms of which the most important later 
outcome was the recitative. 

The following refers to the monodic style itself. 

(547) Giulio Caccini, Nuove Musiche, Florence, 1602, freely transl. 
in Playford, Introduction, London, 1654: 

*I call that the noble manner of singing, which is used without tying a 
mans self to the ordinary measure of time, making many times the value 
of the notes less by half, and sometime more, according to the conceit of 
the words; whence proceeds that kinde of singing with a graceful neglect, 
whereof I have spoken before.' 


(a) Recitative continued to be true to its origins in depending for its 
effect on great freedom of tempo. 



The name recitative* fdry recitative") to the 

most characteristic and Italianate variety, by a 

continue for harpsichord, elf., with string bass. Since are no 

melodic ripieno parts whose movement would impose a of 

measure, the singer is completely at liberty. 

The name recitativo stromentato (Instrumented recitative') 
attached to an alternative variety with parts, the of 

which keeps the music within somewhat narrower bounds, though the 
tempo may still be more flexible than in most forms of measured 

There are other less clearly defined varieties of recitative, or of 
recitative-like melody, some of which (Including, for example., much of 
PurcelPs declamatory writing for the solo voice) are much more 
musically effective If they are taken in a measured tempo with no more 
flexibility than any other expressive melody. 

Only recitativo secco is recitative in the full meaning of the term. 
The less flexible varieties are declamatory music; but recitative in the 
full sense is declamation allied to music, and It Is the declamation, not 
the music, which gives this style its character. 

(547a) Sebastien de Brossard, Dictionaire de Musique, Paris, 1703, 
s.v. 'Largo': 

'[In Italian recitatives] we often do not make the beats very equal, 
because this is a kind of declamation where the Actor ought to follow the 
movement of the passion which Inspires him or which he wants to 
express, rather than that of an equal and regulated measure/ 

Jean- Jacques Rousseau went so far as to consider measured, accom- 
panied recitative a contradiction in terms. He added: 

(548) Jean- Jacques Rousseau, Dictionnaire de musique* Paris, 1768, 
s.v. 'Recitatif Mesure' : 

*One measures the Recitative [only] when the accompaniment . . . 
obliges the Reciter to make his distribution of time-values conform to it/ 

(549) John Hoyle, Dictionary of Music, London, 1770: 

'RECITATIVO: Notwithstanding this sort of composition is noted 
in true time, the performer is at liberty to alter the Bars, or Measure, 
according as his subject requires; hence the Thorough Bass is to observe 
and follow the singer, and not the person that beats time/ 

(b) But though the singer has such wide liberty with the tempo in 
true recitative, this liberty is more subject to conditions than might at 
first appear. 

Declamation is not a mere matter of casual speech enlarged; it has 
a rhythm of its own. There is a symmetry even in its irregularity, and 
unless the singer can sense this symmetry, his liberties will sound not 



expressive, but formless. He hurries on regardless of intelligibility; his 
accompanists, having nothing intelligible to guide them, can only 
follow a fraction of a beat behind; the result is not merely ineffectual 
but untidy. 

A recitative may move extremely fast provided that the singer keeps 
the rhythm of it intelligible and the enunciation clear. Recitative is an 
impassioned art; but it is not an impressionistic art. It has a melodic 
line to be drawn, freely, but with the greatest precision. It must be 
shaped to a pattern which the mind can grasp. An acceleration here is 
matched by a broadening there; an expressive pause in one place by a 
beat brought early in another place. Not only must the accompanists 
be able to keep in touch with all this; they must be in close enough 
sympathy to take much of the initiative. It is, at best, one of the most 
exacting skills; but there is nothing more telling than a harpsichordist 
and string bass player moving as one man, foreseeing every wish of 
their singer, and joining with him. in making every irregularity sound 
not only natural but inevitable. 

Suppose the singer has an up-beat phrase preceded by a little rest on 
the beat itself. He takes the phrase more rapidly than it is written: an 
excellent and dramatic effect. But he must not also take it early, by 
suppressing the rest: this does not sound dramatic, but merely hurried. 
On the contrary, he should take it late, by prolonging the rest, which 
restores the balance. There is now, however, a danger of sounding too 
cumbersome and literal. This is avoided if the accompanists, antici- 
pating his intention, bring the beat itself early, by playing their bass 
note and chord a little ahead of time. This allows the singer to press on 
without losing touch with the beat; for it is then the beat itself which 
presses on. There is a sense of urgency, but not a sense of being out of 
time. This urgency is balanced in its turn by an impression of breadth 
and power if the singer can draw out some corresponding phrase 
expressively. In particular, he will linger over his appoggiaturas as 
though he relished them and could hardly let them go. In these and 
other ways an overall proportion and relatedness can be built up out 
of the irregularities themselves. 

It should be remembered that recitative has not necessarily to be 
forceful. It may also be quietly eloquent and contemplative, as so often 
in J. S. Bach. In these cases, the tempo, though still elastic, may need 
much less irregularity. The style, as always in recitative, should remain 
declamatory; but this is often more a matter of dramatic articulation 
than of irregular tempo. Crisp enunciation supported by warm, can- 
tabile sonority, the necessary declamatory ingredient being supplied by 
the sharpness of the consonants: this is a very satisfactory recipe for 
many kinds of recitative, even including some of the most theatrical. 



See also Ch. XXXII, 2 above, for C. P. E. dis- 

cussion of this problem from the accompanist's of view. 


One of our most harmful reactions over-romanticising 

music has been the sewing-machine rhythm. No music, not music 
based mainly on sequences, will stand a completely tempo. Most 
baroque music needs considerable flexibility. See 1 above for general 
quotations on this point; those which follow are concerned with its 
application in music other than the special varieties already considered. 


A change of tempo between sections or passages may sometimes be 
marked by a change of signature or by a time-word, but is more 
frequently left to the responsibility of the performer. 

(550) Girolamo Frescobaldi, Toccata^ Rome, 1614, preface, 7; 

4 When you find any passage of quavers and semiquavers to be played 
together for both hands, you must not play it too fast. 

[9:] *In the partitas, when you find rapid divisions and expressive 
passages, it will be desirable to play slowly; the same consideration 
applies to the toccatas. Those variations without divisions may be played 
a little more quickly, and it is left to the good taste and fine judgment of 
the performer to regulate the tempo, in which is found the spirit and 
perfection of this style and manner of playing.' 

(551) Thomas Mace, Mustek's Monument, London, 1676, p. 130: 

*[If the music falls into sections or "short sentences", these may be) 
according as they best please your own Fancy, some very Briskly, and 
Courageously, and some again Gently, Lovingly, Tenderly ; and Smoothly! 

(552) C. P. E. Bach, Essay, I, [4th] ed., Leipzig, 1787, III, 28: 

'Passages in music in the major mode which are repeated in the minor 
may be taken somewhat more slowly in this repetition, because of the 


Apart from making sectional changes in the tempo, we may condense 
or stretch it in passing. 

There are two ways of doing this : by borrowing time, and by stealing 
it. In the first case restitution for any extra time taken is made promptly 
enough for the mind to accept the underlying tempo as undisturbed: 
Le. within the bar, or at most within a few bars. In the second case, 
restitution is not made at all: the underlying tempo may be resumed, 
but without any condensing to compensate for the previous stretching. 



The name tempo rubato has been used for both these situations, 
though Its literal meaning, "stolen time\ is strictly applicable only to 
the second. 


(a) We may first take certain technical meanings which tempo rubato 
had for baroque writers, but has not for ourselves. These meanings 
centred round the displacement of rhythm, or sometimes merely of 
accent, within an underlying tempo which is not disturbed. The follow- 
ing Is really a form of that expressive inequalisation of rhythms written 
equally which was always open to the performer in appropriate situ- 
ations: see Ch. XLII, 6 below, and Part III, Rhythm, generally. 
Caccini's notation of his explanation is no doubt approximate; the 
actual rhythm might be very flexible. But the bass would not be involved 
in the disturbance; it would continue with a steady pulse. 

Ex. 192. G. Caccini, Nuove Musiche, 1602, tempo rubato in the 

old sense: 

written .performed approx. 

The following is a late reference to the same idiom, and to another 
in which the mere displacement of the accent serves to suggest an 
ambiguity of time. 

(553) Daniel Gottlob Turk, Klavierschule, Leipzig and Halle, 1789, [1804], p. 40: 

6 Tempo rubato, or robbato, signifies a stolen, or robbed Time, the appli- 
cation of which is likewise left to the judgment of the Performer. These 
words have several significations. Commonly they signify a manner of 
shortening, and lengthening Notes; that is to say, a part is taken from the 
length of one Note and given to the other ... by an anticipation [or] 
by a retardation. Beside this signification, it is understood by Tempo 
rubato, that the Accent is put upon the inferior [off-beat] Notes instead 
of the superior [on-beat] ones.' 

Ex. 193. D. G. Turk, he. cit., tempo rubato in old senses: 


The following is yet another of to we 

should not now apply the term. 

(554) C. P. E Essay, I [4th] ed., 1781 III, 28: 

'[This brings us to] the Tempo mbato. Its indication is simply the 

of more or less notes than are permitted in die measure of the bar. A 
whole bar, part of one or several bars may be distorted, so to speak, in 
this manner. The hardest but most important matter is to give ail notes 
of the same value the same duration. When the performance is so man- 
aged that one hand seems to play across the measure and the other 
strictly with it, the performer may be said to be doing all that is required 
of him. It is only occasionally that all the parts come together. ... A 
master of Tempo [rubato] need not be confined by the numbers [which 
divide notes into groups of] 5, 7, 11, etc. He sometimes adds or omits 
notes to taste . . . but always with fitting freedom.' 

(b) The next quotation describes the normal form of borrowed time 
to which we commonly apply the term tempo rubato. 

(555) Leopold Mozart, Violinschule, Augsburg, 1756, XII, 20: 

'When it is a true virtuoso worthy to be so called whom you are 
accompanying, you must not let yourself be deceived by the ornamenting 
and drawing out of the notes, which he knows so well how to shape 
cleverly and expressively, into delaying or hurrying, but must go on 
playing throughout in the same manner of movement, or the effect 
which the soloist wished to build up would be destroyed by the accom- 

Such descriptions of a timeless melody superimposed on a timed bass 
remind us of Chopin's famous explanation of his own ornamental 

(556) Fryderyk Franciszek Chopin, cited by E. Dannreuther, 
Musical Ornamentation, II, London [1895], p. 161: 

'The singing hand may deviate, the accompanist must keep time 
.... The graces are part of the text, and therefore part of the 
time. . . . Fancy a tree with its branches swayed by the wind; the stem 
represents the steady time, the moving leaves are the melodic inflections. 
This is what is meant by Tempo and Tempo rubato! 

But it is not all that is meant. As a corrective to undisciplined rubato 
in Chopin and elsewhere, the above passage is invaluable. The bass can 
far more often be kept perfectly steady than the undisciplined per- 
former realises; but it must be admitted that there are also many 
passages in which the bass is inevitably involved. In such cases, if 
restitution is made either within the bar or shortly afterwards, the mind 
accepts the underlying tempo as undisturbed. 

In a few exceptional passages, the structure lends itself to much 



longer accelerations compensated by subsequent de-accelerations (the 
opposite effect must be very rare, since normal ritardandos are not 
subsequently compensated). 

(557) C. P. E. Bach, Essay, I, Berlin, 1753, III, 28: 

'[Certain sequential passages] can be effectively performed by accelerat- 
ing gradually and gently, and retarding immediately afterwards.* 

All this still lies within the conception of time borrowed only to be 
subsequently restored. It is an important conception, and Leopold 
Mozart's advice at (555) above must be taken seriously. 


But time may also be stolen, with no intention of subsequently 
restoring it. This, which is now the most popular usage of the term 
tempo rubato, is, if not more important than the foregoing, at least 
more conspicuous. 

The following passages refer both to borrowed and to stolen time, 
but perhaps more particularly to the latter. It will be appreciated that 
the distinction between them is by no means so sharp in practice as it 
has been made to seem here for the sake of clarity. It is, however, a 
valuable one to bear in mind. No tempo rubato should be without shape 
and reason; and it is not at all a bad guiding principle to make restitu- 
tion where restitution can be made, but to have no hesitation over 
stealing time, where it cannot. The main point is to realise that tempo 
rubato in both these normal modern senses has a perfectly ordinary 
place in early music, in spite of the scarcity there of written indications 
for it. 

(558) Girolamo Frescobaldi, Toccatas, Rome, 1614, preface, 1: 

'First, this kind of style must not be dominated by tempo. We see the 
same thing done in modern madrigals, which, in spite of their difficulties, 
are made easier to sing, by means of the flexibility of the time, which is 
beaten now slowly, now quickly, and even held in the air, to match the 
expression of the music, or the sense of the words. 

'[6: With trills in one hand and figuration in the other] let the passage 
flow less quickly, and with expression.' 

(559) Thomas Mace, Mustek's Monument, London, 1676, p. 81 : 

'[Beginners must learn strict time; but] when we come to be Masters, 
so that we can command all manner of Time, at our own Pleasures*, we Then 
take Liberty, (and very often, for Humour [i.e. "mood", not "wit"], and 
good Adornment-sake, in certain Places), to Break Time; sometimes Faster, 
and sometimes Slower, as we perceive, the Nature of the Thing Requires! 



(560) C P. E. I, 1753, III, 8: 

* Certain deliberate disturbances of the arc 

[28 :] "Certain notes and should be 
length for reasons of expression.* 

(561) Daniel Gottlob Turk, and 1789, [1804], p. 40: 

s [In addition to the special forms, ordinary] delaying, or the 

time designedly, is likewise signified by Tempo 


RaEentandos are a case of stolen time, but so important that they 
need separate mention. 

Cadences are extremely numerous in baroque music, and it is the 
nature of a cadence to be acknowledged by a rallentando. If we ack- 
nowledged them all, however, the music would fall to pieces. We have 
to distinguish cadences which pass on at once from cadences which 
bring some portion of the music to an end, or perhaps merely some 
thought. However slight the recognition which we grant the latter, they 
must not be completely over-ridden. A baroque movement taken with 
no internal rallentandos, and then jerked to an abrupt halt by putting 
on all the brakes within a bar or two of the finish, sounds ruthlessly 
insensitive and rigid: like the sewing-machine rhythm referred to in 5 
above, this is part of a puritanical reaction which is not supported by 
the contemporary evidence. 

The proper moment at which to start a rallentando is often suggested 
by the harmony, where it first gives unmistakable warning of the 
approaching close. Most long rallentandos are shallow, most steep 
rallentandos are short; but they must all be graded in a shapely pro- 
portion. No attempt should be made to conceal a weak cadence by 
unnaturally curtailing its rallentando: something may perhaps be done 
by added ornamentation; but in any case it is better to accept the 
inevitable with a show of conviction. The need for rallentandos should 
not, on the other hand, be exaggerated. It is a matter partly of temper- 
ament, but chiefly of a sensitive response to hints offered by the music. 

(562) Girolamo Frescobaldi, Toccatas, Rome, 1614, preface, 5: 

'The closes, though notated as rapid, need to be played in a very broad 
manner; and the nearer you come to the conclusion of the passage or 
close, the more you should hold back the time.' 

(563) C. P. E. Bach, Essay, I, Berlin, 1753, III, 28: 

'In general, ritenutos are better suited to slow or comparatively moderate 
tempos than to very rapid ones. 



[Ed. of 1787 adds:] 'In expressive playing, the performer should avoid 
numerous and exaggerated ritcnutos, which are apt to cause the tempo to 
drag. The expression itself tends to bring this mistake about. In spite of 
beautiful details, the attempt should be made to hold the tempo at the 
end of a piece just as it was at the start, which is a very difficult achieve- 

[II, 1762, XXIX, 20:] 'In slow or moderate tempos, pauses are usually 
prolonged beyond their strict length . . . this applies to [ornamental] 
pauses, closes, etc., as well as to [plain] pauses. It is usual to draw out 
somewhat and depart to some extent from the strict measure of the 
bar . . . the passage [thus] acquires an impressiveness which makes it 
stand out. 

| XXXI, 3:] 'When the principal part draws out before coming to an 
[ornamented] pause, [the accompanist must do the same].' 

The words 'drag' for rallentando and 'away' for a tempo appear in 
early seventeenth-century English MSS., and printed in Thomas Mace 
(Mustek's Monument, London, 1676). 

4 Adagio* or 'grave' near the end of a baroque movement appears 
more likely to stand for molto rallentando (i.e. more drawing out than 
would in any case be assumed without indication) than for meno 
mosso or in tempo ritenuto (at a slower but constant speed): e.g. J. S. 
Bach, G major Prelude and Fugue for organ (E.G. 38, p. 12); B minor 
organ Fugue XIX (B.G. 38, p. 125); C minor organ Fugue XV (B.G. 38, 
p. 105), where the last section is Adagio but has 'adagio' again over 
the last half-bar, perhaps in the sense of piu rallentando. This would 
be not unlike the well-established use of p . . . mf . . ./for crescendo 
and of/. . . mf, . ./> for diminuendo, for which see Ch. XLIX below. 

Ex. 194. F. W. Marpurg, Clavier stucke, Berlin, III, 1763, p. 19, 
rallentando suggested by the word adagio : 

Allegro co 

Adagio piano 


Angelo Berardi, Ragionamenti, Bologna, 1681, p. 136: 

'The representative style . . . consists in this alone, that singing one 
speaks, and speaking one sings/ 







Rhythm is second only to tempo in importance, and like tempo, 
depends eventually on the invisible clock which we carry inside us. 
But unlike tempo, it is sufficiently involved in the structure of the 
music to have an outward shape which can be, and since the middle 
ages normally has been notated. 

The notation of rhythm, however, is necessarily mathematical, 
because our notated time-values go by fixed multiples. The inner 
vitality of rhythm is not mathematical; it is a variable reacting with 
other variables to embody the mood of a given performance. Every 
performer modifies the written rhythms to some extent, sharpening the 
dots in march-like music, softening them for lilting music, and so forth. 
Early performers carried these modifications to extremes some of which 
could at least approximately have been shown in notation. Since they 
were not, they were evidently left to the performer not from necessity 
but from policy: the usual early policy of making nothing in music 
rigid which can be left fluid and spontaneous. That fact is the key to 
the problem of rhythm in early music. 


Unmeasured preludes (see Ch. XXIV, 3, and Ch. XL, 2 above) 
and more or less improvised arpeggiation in figured bass accompani- 
ment (Ch. XXX ? 3 above) or in solos (especially for the violin) written 
as chords but meant to be broken into figurate rhythms at will (Ch. 
XXIV, 3b) : these are among the rare examples of complete liberty of 
rhythm remaining in the baroque period. 


Liberty of rhythm within established conventions may be seen in 
the modification of certain kinds of dotted rhythm (Chs. XLIII-XLIV 



below); and the expressive inequalisation of certain equally written 
rhythms (Ch. XLII below). 


(a) Certain difficulties In notating combinations of duple and triple 
rhythms were habitually evaded, not with any expressive intention, but 
as a mere matter of convenience: the performer is not here being 
trusted with a choice; he merely has to know what the loose notation 
means (see Ch. XLVI below). 

(b) One particularly unpredictable convention concerning the placing 
of the bass note at cadences in recitative (luckily it is known to most 
modern musicians) will be described in Ch. XLII, 4 below. 





In proportional notation, the factors governing the time-value of a 
note, Le. its written rhythm, may include: 

(a) The 'time-signature' (see Ch. XXXVIII above). 

(b) The nominal value of the adjacent notes. 

(c) The form of the note: round, square or oblong; with or without 
upward or downward tail or tails; etc.; also its grouping with other 
notes in ligatures, for which see App. I. 

(d) The coloration: whether open red (or filled-in black) open 
black (== white), etc. 

(e) A dot of augmentation (see Ch. XLIII below). 

(f ) A dot of division, grouping notes for purposes of proportional 

SeeThurston Dart (Grove's Dictionary, London, 1954, s.v. 'Notation') 
for a brief account of the above. 



Of these methods, (b) and (f) were out of use in baroque music, but 
(a), (c), (d) and (e) were in varying degrees retained. 

The use of connecting beams between groups of quavers, semi- 
quavers, etc., in place of strokes to each separate tail, became established 
during the seventeenth century. 

These beams often extended across more notes than is now custo- 
mary, or is very convenient to the eye. Moreover, the groups were not 
necessarily made up in units of a beat, or simple multiples or divisions 
of a beat; the number of notes included may be quite arbitrary, or 
governed more by calligraphy than by ease of reading. But otherwise 
they present no difficulty. 

There may occasionally be a connection intended between the 
grouping by beams and the phrasing; but not ordinarily. Different 
contemporary manuscripts and editions of the same work are quite 
inconsistent in their arrangement of the beams. In particular, groupings 
by pairs and by fours seem indifferently interchangeable. On the 















































- - 

O O O 


O O O O O 






Q^,.., A 







: a 
s u 







occasions a musical significance is is 

easy to see; elsewhere, the is to he for purposes of 


Notice the decorative forms of quaver and semiquaver, which 
rather as if they had one more tail than is actually the case. White 
black forms of the same shape are late of coloration : e.g. the 

black minim (i.e. the crotchet) is worth half the white minim; the 
semlbreve is worth half the white semibreve, and may still be encoun- 
tered in seventeenth-century triple-time. White quavers (equal to 
crotchets) are found in Couperin. 

The longest notes went steadily Into oblivion, the breve being the 
longest In normal baroque usage, as the semlbreve Is In modern usage. 

(564) Christopher Simpson, Compendium, London, 1665, ed. of 
1732, p. 13: 

'The Large and Long are now of little use, being too long for any Voice, 
or Instrument (the Organ excepted) to hold out to their full length. But 
their Rests are still in frequent use, especially in grave Music, and Songs of 
many Parts. 

'You will say, If those Notes you named be too long for the Voice to 
hold out, to what purpose were they used formerly? To which I answer; 
they were used in Tripla Time and in a quick Measure; quicker (perhaps) 
than we now make our Semibreve and Minum. For, as after-times added 
new Notes, so they (still) put back the former into something a slower 

See Ch. XXXV, 2 (b), above for the desirability of reducing the note- 
values in modern editions to protect the performer from being deceived 
into taking much too slow a tempo in music where the discrepancy 
between the apparent note-values in the original and their real speed Is 
very wide. 

The dot of augmentation, which now lengthens the note after which 
it is placed by one-half, had a more variable influence in baroque 
notation: see Chs. XLIII and XLVI below. 


(a) The last note of a piece or section, in all parts, may be written 
with a breve as a conventional indication to hold on to a natural but 
undetermined length much Hke the pause sign often set there nowa- 
days. That this is a mere convention is shown by the fact that some 
parts may be given a breve against a semibreve in other parts; and the 
same part may have different lengths notated in different contemporary 
copies of the same work. It is not intended that the full length should 
be held where the music does not require it; and sometimes, particu- 
larly at repeats, it is most undesirable to do so. 

(b) At phrase-endings other than the end of a piece or section, the 



same convention applies, and it may be desirable to take the last note 
off very much shorter than it is written, in order to avoid fouling the 
next change of harmony or covering the next entry. Thus the autograph 
score of Henry PurcelFs Golden Sonata (No. 9, 2nd set of ten trio 
sonatas, London, 1697) shows at bar six of the fourth movement a 
minim at the phrase end in the string bass-part; but the printed editions 
(just possibly benefiting from a lost intermediate copy corrected by 
Purcell himself, as Walter Emery shows in Two Sonatas by H. Purcell, 
ed. Robert Donington and Walter Emery, London, 1959) give a crotchet. 
The correction makes sure of what Purcell originally assumed that the 
performer would do of his own accord ; but in countless other cases, 
both in Purcell and throughout baroque music, the notation was not 
corrected, and it was left to the performer to avoid an unmeaning 
dissonance or an unnecessary obscurity by adjusting the note-value 
according to the sense of the music. 

Even where no actual dissonance or gross obscurity results from 
holding the literal value, the texture can often be lightened to very 
good effect by not doing so. Cases abound in which one part has a 
longer note-value written than another part which moves with it: there 
is no musical reason why they should differ, and we can again confirm 
that they were not meant to, from the casual discrepancies in different 
contemporary copies of the same works. In polyphonic motets, madri- 
gals and fantasies this point will repay the most careful attention, and 
there is other music in which it is well worth watching. 

(564a) Daniel Gottlob Turk, Pflichten eines Organisten, Halle, 1787: 

'There are even occasions where one must not hold on to the full 
notated value of the note; if, for instance a long note [is notated] after a 
run in quick notes and is followed by a rest, the last note is generally cut 


(565) J. A. Scheibe in F. W. Marpurg's Kritische Briefe, Berlin, for 
1760-62, p. 352, letter 109: 

'Further, some composers are in the habit of anticipating the penulti- 
mate note of the cadence, namely the dominant, without inserting a rest/ 

Ex. 195. Scheibe, loc. cit. 9 naiswritten basses in recitative (corrected 
notation mine, following Scheibe's instructions. But in performance 
the two chords after the voice has stopped can be taken a little late, 
or at least not hurried). 


should be written 


Dotted Notes 


(a) The dot which In our present notation prolongs the value of the 
notes before It by one half Is the same dot which was called the dot of 
augmentation In proportional notation (where It had the same function, 
as opposed to the dot of division, whose function was to separate notes 
for purposes of proportional calculation). 

(b) But in baroque notation, though this prolongation by one half 
was the standard meaning of the dot of augmentation. It was also used 
to mean any convenient prolongation, whether by less than one half, 
or by more. Thus for example: 

TABLE XI: the variable baroque dot 

,,' J- J may equal a J 

;lJ. J33ny l ,dJ_JJ33 

(d) <p * ' rn.iv equal J-. J/ 


(a) The great majority of baroque dotted notes occur in the ordinary 
course of melody and ^are of the ordinary standard length subject only 
to the slight flexibility of ordinary expression. 

(b) But when the dotted notes (i) are persistent enough to dominate 
the rhythm; or (ii) form a distinct rhythmic figure or formula; or more 
generally (iii), would sound sluggish if taken literally: then it was the 
convention to crispen them (Fr. pointer) by lengthening the dot, thereby 
delaying and shortening the note after the dot. This is often called 
'double-dotting'; but since the extent is variable, a better term is 

(a) The lengthening of the dot may vary from the least perceptible 
to perhaps as much as 'quadruple-dotting'. The extent can be quite 



free and unmathematical, except when other parts move across the 
dot; then it Is better to prolong it, if at all in mathematical proportion. 

TABLE XII: dots (a) unprolonged, or 
(b), (c), mathematically prolonged 

or [!>} as 

or (c) as 

(b) A dot followed by a pair or more of short notes may be similarly 
prolonged, and these short notes delayed and shortened: thus Ex. b 
Table XI in (1) b above may be performed as Ex. c. 

(c) When a dotted figure is inverted, the short note or notes coming 
before the dotted note, some authorities recommend a similar sharpen- 
ing of the rhythm, others a milder lilt: see (566) in 4 below. 

(d) For dotted notes against triplet rhythm, see Ch. XLVI below. 


The variability of the dot was taken for granted in late renaissance 
and baroque notation (see Ch. XLV, 5 below for a mid-sixteenth 
century example) and received theoretical attention only at the end of 
the baroque period, when such liberties of notation began to seem 
no longer quite so obvious. 

(566) Joachim Quantz, Essay, Berlin, 1752, V, 21 : 

*In quavers, semiquavers and demi-semiquavers, the standard rule is 
altered, for the sake of the liveliness they ought to express . . . whether in 
slow or quick tempo. 

*. . . it is not possible to determine exactly tke time of the little note 
which follows the dot [but Quantz shows it as follows]. 5 

Ex. 196. Quantz, he. d/., suggested lengths for 'double-dotting': 

[LT nr 

[23 :] 'where the dot comes after the second note [the treatment is the 
same] for the length of the dot and the first note; only the order is re- 
versed [as at (d)]. The same treatment is used with the two little notes at 
[(e) and (f ) ;] the shorter we perform the first [note at (d)] the livelier and 
bolder the expression; and on the contrary the longer we make the dots 
[at (e) and (f )] the more flattering and agreeable the expression/ 


(567) C. P. E. Essay* I 1753, 111, 23: 

The short notes following dotted notes are always 
than their written length. It is therefore to set or 

[of articulation] over them. {The short notes at Ex. (a), (b), (c) and (d) 
would all be taken at the same speed i.e. as dcmi-semiquavers. The short 
notes of the two parts at Ex. (e) would be taken together. It may some- 
times happen, especially at a quick tempo, that the note following the dot 
is best given its written value, to facilitate the movement of the 
parts as at Ex. (f).] 

[Ed. of 1787:] 'When four or more short notes follow a dot, they are 
rapidly disposed of, being so numerous. This also applies to [Ex. (h) to 
(1)], and when the tempo is not too slow, to [Ex. (in) to (q)]. 

[24:] *The first notes of [a Scotch snap when] slurred, are not performed 
too rapidly, in a moderate or slow tempo; for that would leave too long 
a space after them with nothing happening. The first note is stressed with 
a gentle emphasis, but not by a sharp attack or by being prematurely left 
[the opposite of Quantz' advice at (566) above]. 

[II, 1762, XXIX, 15:] 'Since a proper accuracy is often lacking in the 
notation of dotted notes, a general rule of performance has become 
established which nevertheless shows many exceptions. According to this 
rule, the notes following the dot are to be performed with the greatest 
rapidity, and this is frequently the case. But sometimes notes in the 
remaining parts, against which these [dotted notes] have to come in, are 
so distributed that a modification of the rale is needed. Or again, a feeling 
of smoothness which would be disturbed by the inevitably challenging 
effect of dotted notes compels the performer to shorten the dotted note 
slightly. Thus if only one kind of interpretation is used as a starting point 
for performance, the other kinds will be lacking.' 

Ex. 197. C. P. E. Bach, op. cit., I, III, 23, dots of various lengths: 


(568) Leopold Mozart, Violinschule, Augsburg, 1756, I, iii, 11: 

'In slow pieces, there are certain passages where the dot should be held 
rather longer than the [strict] rule requires, to prevent the performance 



from sounding too sleepy ... the time taken by this prolongation is so 
to speak robbed from die note which follows the dot/ 

Ex. 198. Jean- Jacques Rousseau, Dictionnaire de Musique, Paris, 
1768, s.v. 4 Cadence\ 'over-dotting' on trill (interpretation approxi- 
mate only) : 


(569) Jofaann Friedrich Reichhardt, tJber die Pflichten des Ripien- 

ViolMsten, Berlin and Leipzig, 1776, Sect, ii: 

'When [notes] are dotted one after the other, the shorter notes should 
be taken as short as possible to give more emphasis to the longer/ 

(570) D. G. Turk, Klavierschule, Leipzig and Halle, 1789, 48, p. 363: 

'Figures in which the first note is short and the second is dotted are in 
all cases slurred smoothly and flattered. Admittedly the first note takes the 
accent, but the stress should be very mild. 

'Note: it used to be the custom to give the first note of such a figure a 
greatly shortened length/ 


Dots followed or preceded by a pair of notes are normally slurred. 

Dots followed (but not dots preceded) by a short note are very often 
taken in whole or part as silence of articulation. So are dots followed 
by more than two short notes, these notes themselves being sometimes 
slurred, more commonly detached. 

(571) Joachim Quantz, Essay, Berlin, 1752, XVII, vii, 58: 

'The dotted note must be accented and the bow must be stopped (Ger. 
abgesetzet, Fr. dltachf) for the value of the dot/ 

(572) C. P. E. Bach, Essay, I, Berlin, 1753, III, 23: 

'Dots after long notes, or after short notes at a slow tempo, and dots 
occurring singly, are all held on [Ex. (g) above]. But at a fast tempo, 
continuing successions of dots are [often taken as rests] in spite of the con- 
trary appearance of the notation. This contradiction should be avoided by 
a more exact notation. In the absence of that, however, the feeling of the 
music will throw light [on the proper way of performing it]/ 

(573) Leopold Mozart, Violinschule, Augsburg, 1756, 1, iii, 9: 

'In slow pieces ... the dot has to be joined to its note with a diminu- 
endo [and slurred to the subsequent short note]. 

[10:] *In fast pieces the bow is lifted at every dot; hence each note is 
detached from die other, and performed in springing style. 



*[IV, 13: After dots] demi-scniiquavers arc very late the 

next note is played Immediately afterwards a of 

4 [iS: When the dotted rhythm is reversed J the {if. the 

and the subsequent dotted note] are always slurred in 

one bow-stroke. However, the dotted note should not be to die 

away too quickly, nor should it be accented, but held quite smoothly/ 

(574) J. C. F. Rellstab, Anleitung, Berlin, 17%, p. 12: 

'Dotted notes [written] as at [Ex. 198 (a)] are played as at {Ex, 198 (b)]. 
In most cases, the dot is regarded as a rest, and the last note taken shorter 
than its strict value. [In sostenuto passages] the dot is sustained con- 
scientiously, although the last note is still played shorter than its strict value. 
Yet at other times, the short note must be kept at its strict value because 
of the speed or the synchronisation [with other moving parts], A tender 
expression, or longer note values, again make exceptions here; the follow- 
ing passage [Ex. 198 (c)] is played with the notes fully sustained, and with 
the shorter notes taken at their strict value/ 

Ex. 198a. Rellstab, loc, ci/., dotted notes (a), (b), articulated, (c) 

* (a) *c (b) etc 

J. J * ' CU 



(a) Since the double-dot as a symbol of notation existed for several 
generations before being brought into general use in the later eighteenth 
century, the variable use of the single dot was evidently preferred; the 
transition begins to appear in the following two quotations. 

(575) Leopold Mozart, Viotinschule, Augsburg, 1756, 1, iii, 11 : 

It would be very good to decide and set down rightly this prolonga- 
tion of the dot. I at least have often done so, and made clear my intended 
performance by writing a double dot followed by a shortened note/ 

(576) F. W. Marpurg, Anleitung zum Clavierspielen, Berlin, 1755, I, 
x, p. 13, Anmerkung: 

If the dot is to be very sharp, i.e. if the succeeding short note is taken 
with more vivacity than its [written] value prescribes: then one should in 
the first place write two dots one after the other, and in the second place 
shorten die succeeding note by half/ 

(b) In Ex. 199 (a) below Couperin has made certain of the 'double- 
dotting' by adding an extra tail to the short notes; yet he still does not 



use the double-dot, either because he was not aware of It, or because 
he saw no reason for It, being accustomed to using the dot for any 
required lengthening at will e.g. as at Ex. 199 (b) below. The reason 
for the extra tall is that without It the dotted notes, so far from being 
"double-dotted", might have been reduced to triplet rhythm, to fit the 
triplets which come against them, under the convention discussed In 
Ch. XLVI, below. 

Ex. 199. (a) Couperin, UApotheose de Core///, Mt. VI, dot used 
instead of double-dot; (b) Couperin, XIV Concert, Prelude, bar 1, 

dot used at convenience: 

(c) The following is a rare exception: i.e. an accurate notation of 
'double-dotting', complete with written-out silence of articulation. The 
explanation is obvious: it is an Italian composition in French overture 
style, which would not at that date be familiar to Italian performers. 

Ex. 200. Arcangelo Coreili, Op. VI, No. 3, publ posthumously, 
Amsterdam [1714] opening Largo: 

(d) At Ex. 206 in Ch. XLV, 8 below, Handel has notated equal notes 
at the opening statement but unequal notes at its recapitulation. The 
intention is quite certainly that the inequality should be present from 
the start, since there can be no conceivable musical justification for 
performing the two passages in different rhythms from one another. 
Such cases abound. 


The convention of "over-dotting' certainly survived into the 
generation of Beethoven see (574) in 5 above, and probably lost force 


gradually; it still survives to the to 

musician will sharpen a crisp thai he is 

so. It Is a natural instinct. It wants carrying in 

however, than Instinct alone will carry it; and it is, in fact, one of the 
reforms which make most difference to the vitality of music. 

Passages which have seemed inexplicably into life and 

energy when the "over-dotting" has to with 




Other Sharpened Rhythms 


(a) French Overture style is a convenient name for one of the most 
characteristically French contributions to baroque music. This is a 
certain idiom of composition which originated in the slow introductory 
movement of the French Overture itself, and spread to many other 
contexts and countries. It can be recognised by its special feature, 
which is a remarkable contrast between a ground-swell of majestic 
rhythm and a surface turbulence of dotted notes and rushing scale- 

This contrast is intended to be exaggerated in performance. The dots 
are strongly 'over-dotted'; and they are taken wholly or largely as 
silences of articulation. The scale-passages, some three to seven notes 
long, are left as late as possible, and then taken very rapidly, but 

When the scale-passage is written as the up-beat after a dotted note, 
it is a particular instance of ordinary 'over-dotting'. When it follows 
the beat after a short rest, it is this rest which is prolonged, and the 
result may be regarded as an extension of the 'over-dotting' principle. 

The entire effect is one of ferocious flexibility on a foundation of 
strength and poise. In a literal performance, this effect is lost, and the 
music merely sounds heavy. French Overture style depends even more 
than most baroque music on the invigoration given it by conventional 
departures from the written rhythm. 

(577) Joachim Quantz, Essay, Berlin, 1752, XVII, ii, 16: 

'When some demi-semiquavers come after a long note and a short 
rest, they should always be taken very quicldy, whether in Adagio or in 
Allegro. Before playing them, you should wait until the very end of the 
time allotted to them, so as to avoid getting out of \i,e. ahead of] time/ 

Ex. 201. Quanta, op, cit., runs in French Overture style (the 
interpretation is mine): 

[Quantz, Ch. XII, 24:] "The majestic style is by 

during which other parts have rapid by 

These latter must be enforced with power attack. The dot is 
and the following note thrown quickly away. 

[vii, 58:] "When there are three or more semiquavers a dot, or a 
rest, they should not be given their strict valee, particularly in 
but after waiting until the very end of the time allotted to them, you 
play them as fast as possible, in a style frequently found in Overtures, 
Introductions and Furies. Nevertheless you must give each one of 
quick notes a separate bow, and you can hardly slur anything. 

[iv, 10: The 'cellist] 'should play his dotted notes more gravely and 
weightily than the violins; but when it comes to the semiquavers which 
follow them, he must make them short and sharp [I.e. with the violins] 
whether in slow time or in quick/ 

(b) The following fugue by J. S. Bach is an excellent example of the 
wider influence of the French Overture style, and is given here In 
Arnold Dolmetsch's interpretation (Interpretation* London [1915], p. 
65 and in his fine recording on the clavichord for the Columbia *Forty- 
Eight* Society: a wonderful combination of impetuosity, nobility and 
sheer beauty of tone). 

Ex. 202. J. S. Bach, Fugue in D major, Bk. I of the 'Forty-Bight*, 
interpretation by Arnold Dolmetsch, loc. tit. : 


(a) The following convention will be found of very wide application. 
(578) Joachim Quantz, Essay, Berlin, 1752, XVII 9 ii, 16: 

'If in a slow allabreve or in normal common time there is a semiquaver 
rest on the accented beat followed by dotted notes, you must take the 



rest as if it had a dot attached to it, or [another] rest of half its value, and as 
if the note after it were a deniiseniicjuaver.' 

Ex. 203. Quantz, loc. a'/., shortening the note before a dotted 
rhythm (the interpretation is that given by Arnold Dolmetsch, 
Interpretation, London [1915], p. 58): 

.(a) If as written] 'h] 

'a} [takf n apprvix ] 

In Ex. 203a following, there is no distinction of rhythm intended; 

the conventional y^J^ should be maintained throughout. Cases of 
this kind are exceedingly common. 

Ex. 203a. Handel, Op. I No. 1, 1st movt., Grave, sharpened 
rhythms not all correctly notated : 

Bar 14 

Bar 16 

(b) A detail which falls under the same application of the principle of 
"over-dotting' is the little preliminary note or pair of notes in certain 
dance movements such as allemandes, courantes and ayres. 

Ex. 204 (a) below shows the normal 'over-dotting'. Ex. 204 (b) 
shows the single preliminary note delayed and shortened as if a hypo- 
thetical dotted note, or rather a hypothetical rest, were in front of it. 
(For the sharpening of the dotted rhythm specifically in courantes, 
see (592a) in Ch. XLV, 8 below. For the uneven rhythm in the inter- 
pretation of Ex. 204 (b), see Ch. XLV below.) 

Ex. 204. Matthew Locke, Suite II for four viols, mid- 17th cent., 
(a) Courante, (b) Ayre: 

( a ) written. 

ifitC, ' ^^ ~~-^. 

* [Fast] ' 
( b ) written 

played approx. 



The Almand of PurcelPs 2nd Suite for Harpsichord (G minor) has 
this little preliminary note shown as a semiquaver at the start of the 
first half, and as a quaver at the start of the second half. The foD owing 


Courante has the reverse of this. No distinctions are the 

notation is casual but a semiquaver is evidently 

there being no musical justification for any such inconsistency in the 

performance. These and other similar strongly to confirm 

the interpretation here suggested. It is so often 

inconsistencies of notation which help to show the consistent 

at the back of them. 


The mistake (marked sic where hemi-demi-semiquavers are required) 
in Ex. 202 above has been corrected in the 1946 edition of Dolmetsch's 





Among the expressive liberties taken for granted by early performers 
was the right to modify a four-square rhythm by performing unequally 
certain notes which are written equal. 

(579) Michel de Saint-Lambert, Prlndpes du Clavecin, Paris, 1702, 
p. 25: 

fi [We make certain notes unequal] because the inequality gives them 
more grace. . , . Taste judges of this as it does of tempo.* 


(a) 'Over-dotting' is for exhilaration; but inequality is for per- 
suasiveness. It has no place in essentially four-square music: marches; 
allemandes, perhaps (but opinion differed here); indeed most, though 
not all, common time movements of which the pulse is decidedly four 
in a bar. 

Common time movements of which the pulse is decidedly two in a 
bar, and all varieties of triple time, lend themselves best to inequality. 

(b) Inequality is out of place in mainly leaping melodies; it is highly 
suitable to mainly stepwise melodies. 

(c) Only those notes which fall naturally into pairs are liable to 
inequality. If they are notated slurred in pairs, and are otherwise 
eligible, this is virtually an indication of inequality. They must in any 
case be slurred by pairs in performance if taken unequally. Notes 
slurred more than two at a time do not fall into pairs and are not 
eligible for inequality. 

(d) Notes so slow that inequality would sound sluggish or so fast 
that inequality would sound restless are not eligible for inequality. 

(e) Only the fastest notes appearing in substantial numbers are 
strictly eligible for inequality. Thus in quick time, the presence of semi- 
quavers in substantial numbers prevents inequality, since they are too 
fast themselves yet by their presence they make quavers ineligible. But 
if only a few are present, they can be ignored, and the quavers may 
then become eligible. At a slower time, semiquavers are eligible if 
demi-semiquavers are not present in substantial numbers; etc. See also 
(g) below. 



(f) The ordinarily for are go 
either two or four to a beat such as: 

TABLE XIII: different time-signatures 


3 432 6432 12 9643 

I 222 4444 88888 

Minims Crotchets Quavers 

Crotchets Quavers Semiquavers Demi-semiquavers 

(g) When such notes are mixed with substantial numbers of shorter 
notes not too fast to be eligible for inequality, either the shorter notes 
only, or both at the same time, may be taken unequally, in spite of 
what has been said at (e) above. 

(h) Notes otherwise eligible for inequality, but bearing dots or 
dashes, are specifically precluded from being taken unequally. 

The dot is not a regular baroque sign for staccato, though it some- 
times has that meaning. It is a regular sign for equal notes. 

The dash is one regular baroque sign for staccato ; but since notes all 
staccato cannot be grouped in pairs, it incidentally precludes inequality. 

(i) Notes otherwise eligible for inequality can also be specified as 
equal by such words as: egalement; notes egales; notes martelees; 
detachez; mouvement decide; mouvement marqu6; coups 6gaux; etc. 
See also (583) in 7 below. 

Inequality can be specifically indicated by such words as: inegales; 
notes inegales; lourer; pointer; etc, 


Here the first note is lengthened and the second shortened, usually to 
about triplet rhythm, J^ becoming J 3 J" etc. A modern editor wanting 
to suggest the lilting lourer approximately in notation can often best do 
so by transcribing the entire piece from J into l j or from | into |, etc. 

The actual proportioning, however, is at the performer's discretion. 

Any dotted notes notated, in among the equal notes thus inequalised, 
should be sharply 'over-dotted' to distinguish them from the milder 
inequality. The term pointer is found for this but also in the more 
general sense of 'point the rhythm' in whatever degree. 

Conversely, a passage notated consistently in dotted notes may, as 
an alternative to 'over-dotting' it, be softened to the lilting lower 
rhythm. Many of the long dotted passages found so characteristically 
in Purcell and his contemporaries sound wonderfully well in this softer 
interpretation, though others need the brilliance of the sharper rhythm, 
It is a salutary shock to realise that this choice is freely trusted to the 
performer. We can do which we please. 




Whereas the lilting rhythm helps the music along, the snapped 
rhythm sometimes known as coaler (but this term, literally 'slur', has 
many other meanings) is inclined to hold it back with its limping gait. 
See Cfa. XLIII, 4 above for the various degrees of jerkiness recommended 
by different writers for this rather jolting rhythm. It can be taken 
sharply, like a Scotch snap, but can also be flattered into a gentle 
eloquence better suited, for example, to the typical Purcell passages in 
which it is written out. Here, the first note is shortened. 

A pair of notes slurred without dot can mean the same. 

The snapped rhythm, as opposed to the lilting rhythm more likely 
to be favoured by a baroque performer in the absence of other instruc- 
tions, could be specifically indicated by a slur over pairs of notes and 
a dot over the second. Contrary to the modern significance, this meant: 
lengthen the note bearing the dot over it at the expense of the note 
before it, and thus introduce the snapped rhythm. See (586) in 7 below. 


(580) Fray Tom4s de Santa Maria, Arte de taner fantasia, Valladolid, 
1565, 6 7th condition': 

'Concerning performance in good style, which is our seventh condition, 
it is to be remembered that it is necessary for this to play the crotchets in 
one way, the quavers in three ways. 

The method to be observed in playing crotchets is to linger on the first, 
to hurry on the second, to linger again neither more nor less on the third, 
and to hurry on the fourth. This is done as if the first crotchet were dotted, 
the second crotchet a quaver, and likewise as if the third crotchet had a dot, 
and the fourth crotchet were a quaver, and so on. And care must be taken 
that the hurried crotchet should not be too hurried, but only slightly 
so ... 

'[The first method with quavers is] to linger on the first quaver and 
hurry on the second, to linger again neither more nor less on the third, 
and so on. This is done as if the first quaver were dotted, the second 
quaver a semiquaver, and likewise as if the third quaver had a dot, and 
tie fourth quaver were a semiquaver, and so on. This method is used in 
works which are entirely contrapuntal, and for long and short passages of 
free ornamentation (glosas). 

'In the second method, we hurry on the first quaver, linger on the 
second, hurry neither more nor less on the third, linger on the fourth, and 
so on. This is done as if the first quaver were a semiquaver, the second 
dotted, the third again a semiquaver, the fourth dotted, and so on. . . . 
This method is used in short passages of free ornamentation (glosas) 
which are made in [existing] works (obras) as well as in fantasies. 

'Observe that this method is more graceful than the foregoing. 



In the third method, we hurt}' on three quavers, on the 

then hurry on the next three, linger on the fourth, that we 

must linger long enough for the fifth quaver to arrive in its proper 
At the minim [interval], and so on, so move in of four 

and four. This is done as if the three quavers were semiquavers, the 
fourth quaver dotted. This third method is the most graceful of all, and 
is used for long and short passages of free ornamentation 

4 Care should be taken not to linger too long on the quavers* but Just 
Song enough to point them, since long fingerings are the of 

clumsiness and ugliness in music; and for the same reason, the three 
hurried quavers must not be hurried too much, but only slightly, to 
correspond to the amount of lingering on the fourth.' 

We have here a vivid description of the lilting rhythm (paragraphs 
two and three) and of the snapped rhythm (paragraph four), though 
without the detailed conditions which later writers attached to them; 
and also of a third rhythm (paragraph six) which later writers do not 
seem to describe (but cf. the last of Caccini's modifications at Ex, 205 
(c) in 6 below). 

We may particularly notice the freedom of rhythm described: *the 
hurried crotchet should not be too hurried, but only slightly so*, etc. 
The three semiquavers and dotted quaver in paragraph six do not even 
add up, mathematically, to the minim's length required. The whole 
account shows that mathematical subdivision is not what is intended. 
The inequality is not measured, but proportioned expressively at 

This gives us incidental evidence that the dot was regarded in 1565 
as a sign for unmeasured prolongation, as well as for measured pro- 
longation by one half. The phrase 'somewhat dotted" used by Fresco- 
baldi at (581) in 6 below is further confirmation; and so also is 
Couperin's use of the word 'dotted' in describing inequality at (586) 
in? below. Compare also the following: (580a) G. B. Bovicelli, Regale 9 
Venice, 1594, p. 1 1 : 'the more the first note is lengthened and the second 
shortened, the more graceful is the melody/ 


Ex. 205. Giulio Caccini, Nuove Musiche, Florence, 1602, preface, 
(a) lourer, (b) couler, (c) free rhythmic modifications: 

as written (b) performed, approx. 





00 (c> 

(581) Glrolamo Frescobaldi, Toccatas, Rome, 1614 5 preface, 7: 

'When you find any passage in quavers and semiquavers to be played 
together by both hands, you must not play it too fast; and the hand which 
has semiquavers should make them somewhat dotted; dotting not the 
first but the second, and so on with the others, one without dot, the other 


(58 la) B. de Bacilly, UArt de bien chanter, Paris, 1668, pp. 232-3: 

'Although I say that alternate dots are implicit in divisions (that is to 
say that of two notes one is commonly dotted), it has been thought best 
not to mark them for fear of their being performed by jerks . . . these 
notes must be dotted with such restraint that it is not obvious (except in 
certain pieces which require this manner of [jerky] performance). And 
indeed it is necessary in some passages altogether to avoid dotting/ 

(582) Georg Muffat, Florileglum /, Augsburg, 1695, preface: 

*All die difference [between a slow 2 with two in the bar and an ordi- 
nary C with four in the bar] consists, in that under the last, several quavers 
continued in succession , , J" J* etc., cannot be alternately dotted \. \ 

V V 

J\ j* etc. for elegance in performance, like the others; but should be 
expressly strictly the one equal to the other [because of the four-square 

(582a) Menne Loulie, Elements . . . de musique, Paris, 1696> 
Amsterdam ed. of 1698, p. 38: 

'In each measure but especially triple measure, the half-beats are per- 
formed in two different ways, although they are notated in the same way. 

c l. They are sometimes made equally. This way is called detaching the 
notes, you use it in melodies of which the sounds move by leap, and in all 
kinds of foreign music where you neper inequalise (pointe) them except where 
marked [but this latter remark seems untrue or much exaggerated: see 
8 below]. 

*2. Sometimes the first half-beats are made a little long. This way is 
called Lourer. You use it in melodies of which the sounds move step- 
wise . . . 



"There is further a way, the is 

longer than the second, but [for this to be appropriate] the 
ought to have fin the notation] a {written dot]. 

*This kind is called Piquer, or [no is of the 

which was undoubtedly much less widespread important].' 

(583) Sebastien de Brossard, rfe Paris, 1703, 
5.r. *Artdante': 

'Andante . . . means above all for Basso-continuos, al! the notes 
must be made equal, and the sounds well separated." 

(584) Jacques M. Hotteterre, Principes de la Fluie Traversi&re, Paris, 

1707, p. 24: 

*It will be observed that quavers are not always to be played equally, 
and that in some measures there should be a long one and a short one; 
this is governed by their number. When this is even, the first is to be 
taken long, the second short, and so on. When it is an odd number, the 
opposite arrangement is followed. This is called pointer [NJB. the con- 
fusion of names here contrast (582a) above]. The measures in which this 
is most usual are two in a bar, | and 4.' 

(584a) Michel de Monteclair, Nouvelle Methode, Paris, 1709, p. 15: 

'It is very hard to give general principles as to the equality or inequality 
of notes, because it is the style of the pieces to be sung which decides it; 
however, you should observe that in all varieties of measure the notes of 
which four go to a beat are intended to be uneven, the first a little longer 
than the second/ 

(585) Michel de Monteclair, Petite Methode, Paris [c. 1730], p. 42: 

'The length of the quavers, in the two-time measure marked simply by 
a 2, is unequal; the first quaver lasts almost as long as if it were followed 
by a dot, and the second almost as quickly as a semiquaver [unless marked 
"croches 6gales"]. 

'[The 4 is beaten with two quick beats; die quavers are equal and the 
semiquavers unequal.] 

'[Heavy three time is marked 2, but gay three time is marked 3 or J and 
in this] the quavers are unequal. 

'[Quick three time is marked | and in this] the quavers are equal. The 
semiquavers are unequal. 

[p. 48:] 'The plain C marks the heavy four time measure, and the 
stroked Q marks the quick four time measure. However, when the move- 
ment is quick to the point of tiring the arm, this measure can be beaten 
with two beats, while retaining the equality of the quavers. 

'[p. 50: In 4 as in |] the quavers are unequal. 

'[In there are] two equal beats; there are three equal quavers to each 


(586) Fr. Couperin, Pieces de clavecin, Paris, 1 71 3, table of ornaments : 

4 Slurs, of which the dots indicate that the second note of each beat 
must be stressed/ 

Ex. 205a. Fr. Couperin, loc. c//., notation indicating stress on the 
second note of each pair: i.e. the couler (interpretation by Arnold 
Dolmetsch, Interpretation, London [1915] p. 73): 


(587) Joachim Quantz, Essay, Berlin, 1752, XI, 12: 

'A very necessary remark must be made, with regard to the time 
which is given to each note. A difference must be made in the rendering 
between capital notes, which are also called accented (frappantes), or 
among the Italians good notes, and those which pass, and which some 
foreigners call bad notes. The main notes ought always, if possible, to be 
brought more into relief than those which only pass. According to this 
rule it is necessary in pieces of a moderate speed and even in Adagio for 
the quickest notes to be played with a certain inequality, even though 
they appear at sight to be of the same value; so that at each figure the 
accented notes, to wit the first, third, fifth and seventh, must be pressed 
upon more than those which pass, to wit the second, fourth, sixth and 
eighth, though they must not be sustained as long as if they were dotted. 
By these very quick notes I mean the crotchets in triple-minim time, the 
quavers in triple-crotchet time, the semiquavers in that of quavers, the 
quavers in Allabreve, the semiquavers or the demisemiquavers in four- 
eight time or four time. This no longer occurs, however, as soon as these 
notes are found mixed with figures of notes yet quicker or half as short in 
the same time; for then these latter must be played [with] the first and the 
third of the four notes a little pressed upon, and their tone made a little 
louder than that of the second and fourth notes. But we except from this 
rule in the first place quick passages, in a very quick movement, where 
it is only possible to press and use force on the first note of the four. We 
further except all those passages which singers should render with speed, 
unless they ought to be slurred [in pairs] ; for since each note of this sort 
of passage for the voice ought to be made distinct and marked by a little 
stroke of the chest [ski coup de poittine], inequality has no place there. 
Lastly we should except notes on which there are dashes or dots. It is also 
necessary to make the same exception, when several notes follow at the 
same pitch, or when there is a slur over more notes than two, to wit four, 
six or eight; and lastly with regard to quavers in Gigues. All these notes 
ought to be rendered equally, the one no longer than the other. 

4 [XVII, 2: With semiquavers in a slow movement] give more emphasis, 
both for length and for volume, to the first of a pair than to the second 



[and take a quaver followed by two semiquavers] almost as though 

[XII, 12:] "When the first of four semiquavers is a rest, you must delay 
half as long again as the written value of the rest, because the second note 
[of each pair] should be shorter than the first/ 

(588) C. P. E. Bach, Essay, II, Berlin, 1762, XXIX, 15: 

'The semiquavers [at Ex. 205b] sound feeble in an adagio unless dots 
are placed between them [i.e. unless made unequal]/ 
Ex. 205b. 

(589) Dom Fr. Bedos de Celles, LArt du Facteur d'Orgues [Paris] 
1766-78, p. 600: 

'Even notes of equal [written] value [are] more or less prolonged. 

[p. 601:] 'In all movements in 2, 3 and 4 time, the quavers are 
articulated 2 by 2, and are distinguished as first and second: this distinction 
also sometimes takes place with crotchets . . . 

'This distinction between first and second, can also occur in semiquavers 
of 4 in a moderate movement . . . 

'Almost always the first are longer, and the second shorter; I except 
however movements where they are articulated by 3 and 3, as in the 4 
and |: but in movements where they are articulated by 2 and 2, it is rare 
that they should be equal. 

'This inequality should vary according to the kind of expression of the 
melody; in lively melodies, it should be more marked than in melodies 
which are gracious and of a tender expression, [more marked] in marches 
than in minuets; however, plenty of minuets are found in which the 
inequality is as marked as in a march [Dom Bedos evidently has in mind 
here inequality so extreme as to become equivalent to dotting perhaps 
even to "over-dotting"]/ 

(590) M.-D.-J. Engramelle, La Tonotechnie, Paris, 1775, p. 32: 

'There are cases in which this inequality is one-half, so that you should 
perform the first as if they were dotted quavers, and the second semi- 
quavers; other cases in which the inequality is one-third, as if the first 
were worth two-thirds of a quaver, and the second the remaining third; 
others again in which this inequality is less conspicuous and should be as 
3 to 2, so that the first will be worth three-fifths of a quaver and the 
second two-fifths. 

[p. 33:] 'There are many marches ... in which the inequality between 
the first and the second is as 3 to 1. In certain minuets ... the inequality 
is as 2 to 1 ; finally, in many minuets, the inequality is less conspicuous, 
e.g. as 3 to 2 or as 7 to 5. 



[p. 230:] "Inequalities ... in many places vary in the same piece; it is 
left to fine taste to appreciate these variations of inequality ... a little 
more or less inequality substantially changes the expression of a piece.' 


This is one of the most difficult questions in baroque interpretation. 
Some Frenchmen write almost as if they had a monopoly of in- 

(591) Fr. Couperin, UArt de toucher, Paris, 1716, ed. of 1717, p. 38: 

'There are in my opinion defects in our notation of music comparable 
to those in the writing of our language. We notate otherwise than we 
perform, which is the reason why foreigners perform our music less well 
than we perform theirs. The Italians on the contrary, write music as they 
have conceived it. For example, we perform as dotted a succession of 
quavers moving by step; yet we notate them as equal.' 

In spite of its distinguished author's experience in writing in the 
Italian as well as in the French style, this passage is unquestionably 
misleading. A truer picture emerges from the following. 

(592) Michel Corrette, Methode de la Flute Traversiere, Paris-Lyons, 
[c. 1730], p. 4: 

'The four-time C or (f is much used in Italian music, as in the Alle- 
mande, Adagio, Presto of sonatas and concertos. 

It is necessary to perform the quavers equal and to make unequal 
(pointer) the semiquavers two by two. These are also sometimes per- 
formed equal in Allegros and Prestos of sonatas and concertos. 

'The 2 marks the measure of two beats. This measure is used for 
Rigaudons, Gavottes, Bourrees and Cotillons in French music; the 
Italians never use it. The quavers must be made unequal two by two, that 
is to say make the first long and the second short . . . 

'The 4 or I is the 2-time of the Italians. This measure is often used in the 
Allegros and Prestos of sonatas and concertos. The quavers must be 
performed equal, and the semiquavers made unequal: they are also 
sometimes performed equal in sonatas. 

'The 3 is used for French and Italian Jigs. The quavers are performed 
equal, the first two are very often taken with one stroke of the tongue and 
sometimes three. 

'The I marks a slow movement. This measure is sometimes used for the 

'The crotchets must be made unequal two by two, and sometimes 
performed equal according to the character of the piece . . . 

'The I is used in French music for Passepieds. This measure of three 



beats is very often found in the AiFettioso, Minuets, and Allegros of 

'The quavers are taken equally except for the semiquavers which are 
made unequal. 

*The 4 is a measure of two uneven beats [i.e. containing an uneven 
number of units, namely three crotchets to each beat] ; it is used for the 
Loure in French music. The English compose many Vaudevilles, and 
Country-dances in this measure. . . . These airs ought to be played in a 
noble manner, marking the crotchets well; and making the quavers 
unequal two by two. This measure is very little found in Italian music. 

"The f is found in Italian, German, French and English music, in 4-time 


'The quavers must be performed equal and the semiquavers made 

'The first two quavers and sometimes three are taken with a single 
stroke of the tongue. 

'[Methode, Paris, 1741: In the Courante] the quavers are performed 
equally in Italian music as for instance in the Courante of Corelli's 
Sonata Op. 5 No. 7. But in French music the second quaver in each beat 
is performed more quickly/ 

(592a) Joachim Quantz, Essay, Berlin, 1752, XVII, vii, 58: 

'In that measure [(P two in a bar], as well as in the time of three crotchets 
in a bar, which is used for the Loure, the Saraband, the Courante, and the 
Chaconne, the quavers following the dotted crotchets are to be performed 
not in their strict [written] value but very shortly and sharply. . . . All 
dotted notes are to be performed in the same manner wherever there is 
time to do so/ 

The musical situations in which the possibility of inequality arises are 
so numerous, and the possibilities themselves are so extensive, that a 
modern editor may have great difficulty in reaching suitable decisions. 
The following are comparatively simple cases. The Handel tune appears 
first in mainly equal notation, later in unequal notation. There can be 
no reason other than inadvertence; but though the notation was casual, 
the intention was not, and in spite of appearances the movement is to 
be taken consistently in unequal rhythm. 

Ex. 206. G. F. Handel, Op. I, No. 1, c. 1724, opening slow 

Gravc Bar 12 



In an earlier version of this sonata, there is a slur over the first E, D 
sharp and E in both bar one and bar twelve, the rhythm being even in 



both these bars; the next four notes are unslurred but uneven In both 
these bars. The two versions contain many similar discrepancies, which 
may be readily and valuably compared in the Urtext available as Lea 
Pocket Score No. 70. 

It will be appreciated that slurs are among the details least likely to 
be carefully or consistently notated; and also that a phrase which 
appears equally notated in one contemporary edition will often appear 
unequally notated in another. No reliance can be placed on any of 
this; it is the convention, not the notation, which we have to take into 
account. It is true that the contemporary Interpreters would have used 
various amounts and forms of inequality: so may we; but we must 
use, In general, enough. 

A hint which we can sometimes take from the notation concerns the 
choice between the lilting rhythm and the snapped rhythm. Thus in the 
duet Domine Deus of the Gloria in J. S. Bach's B Minor Mass, the nota- 
tion shows the snapped rhythm at the first appearance of a theme which 
returns in equal notation: the snapped rhythm should be maintained 
throughout. Again, in transcribing his own E major violin concerto 
for the harpsichord, J. S. Bach went to the trouble of changing the 
notation from equal to unequal, also in the snapped rhythm, and 
presumably with the object of securing this rhythm in preference to the 
lilting rhythm which the violin version perhaps most naturally sug- 
gested. It will be seen, Incidentally, that whereas the steps of the melody 
are made unequal, the leaps are left equal: this is quite certainly inten- 
tional (see 2 (b) above). It will be remembered that J. S. Bach was 
notorious for including in his notation many elements of figuration 
which were normally left to the performer in his day: see Ch. VII, 3, 
(112) to (113) above. We can therefore often learn from his written 
text valuable lessons as to what other composers expected from the 
performer. Even J. S. Bach, however, did not by any means write down 
everything he intended and expected. 

Ex. 207. J. S. Bach, written-out couler (a) in the Gloria of the B 
minor Mass, (b) in transcribing the E major Violin Concerto, Movt. 
II, for harpsichord: 

(b) Vn. Concerto: 

Rhythm changed in harpsichord concerto to : 


The generation of Beethoven appears to be the last in which 
contemporary authorities take inequality for granted. 



Triplet Rhythm 


(a) Triplets, quadruplets, quintuplets, sextolets, septuplets, etc., are all 
"proportions' in the old sixteenth-century sense, and by sixteenth 
century musicians, may have been performed literally. But the seven- 
teenth century lost interest in 'proportions', and through the baroque 
period, an appearance of 'two against three' or 'three against four' is 
deceptive. Three against four' as a genuine proportion may recur in the 
galant style, e.g. of G. P. Telemann; by the mid-eighteenth century, it 
is becoming difficult to be sure whether a real 'two against three' is 
intended. But neither of these proportions had an accepted place in 
true baroque rhythm. 

Triplets as such are common. We find them written in the modern 
way as groups of three notes, with or without a 3 to make the grouping 
clear, at least from Petrucci's famous editions at the beginning of the 
sixteenth century. But the duple notes set against them in baroque 
music are not meant for cross rhythms; they are meant to accommodate 
themselves to, or be accommodated by, the rhythm of the triplet. 

The notation is a mixture of half-remembered proportions and loose 
conventions. The convenient flexibility of the dot and the familiar 
practice of inequality were both pressed into service: J3 was softened, 
and J~3 was sharpened, to produce the same result (i.e. J 3 T). 

(593) C. P. E. Bach, Essay, I, Berlin, 1753, III, 27: 

'Now that triplets have come increasingly into use in common or 
4 time, as well as in \ and 4, many pieces have made their appearance 
which could with greater convenience be notated in g 2 , | or |. The per- 
formance of other values against these [triplet] notes is shown [at Ex. 208].' 

Ex. 208. C. P. E. Bach, be. eft., (a) dotted notes compressed and 
(b) equal notes expanded into triplet rhythm when set against 


(b) Quantz, like C. P. E. Bach, treats the standard compression of 
dotted rhythm into triplet rhythm (bringing the little note after the dot 



Into line with the last note of the triplet) as the strict meaning of the 
notation; he treats what to us is the strict meaning of the notation 
(bringing the little note after the last note of the triplet) as a case of 
'over-dotting* no doubt because if the dot Is thus taken at all, It 
should be 'over-dotted' to emphasise the effect. He prefers this to 
the standard interpretation, which is that given by C. P. E. Bach. 
(594) Joachim Quantz, Essay , Berlin, 1752, V, 22: 

'This rule [of "over-dotting' 5 ] ought also to be observed, when one 
of the parts has triplets, against which the other part opposes dotted notes. 
It is then necessary to sound the little note which follows the dot, only 
after the third note of the triplet, and not at the same time with It. Other- 
wise we could confuse them with | or * 8 2 time/ 

But we have seen at (593) above that C. P. E. Bach regarded such 
cases as being actually equivalent to ! 8 2 , | or f time; and since his book 
appeared just the year after Quantz', he may have been refuting him. 
Thus both ways must have been in use by then. 

(c) The following case, which dates from a generation before either 
Quantz or C. P. E. Bach, is important because it shows how naturally 
the convention of accommodating duple to triplet rhythm arose. The 
movement starts with a correct use of proportional notation, and with 
the correct time-signatures to denote it: Ex. 209 (a). But by Ex. 209 
(b) the notation has fallen back on mere common sense, and is no 
longer correct at all. The interpretation, however, is made quite certain 
by the imitation of rhythm between treble and bass : it is as at Ex. 
209 (c). 

Ex. 209. Arcangelo Corelli, Op. V, No. 3, Rome, 1700, Allegro, 
(a) correct proportional notation degenerating into (b) loose con- 
ventional notation meant for (c) : 

'(a) -Allegro ctc 







1 ^_ 

* pT 

-. .M. 







(d) The next is a case of perfectly correct proportional notation, 
interesting because its composer had evidently forgotten why it was 



^customary to write thus'. Dolmetsch (Interpretation, London, [1915], 
p. 68) also mistook it for "incorrect writing' conventionally accepted, 
as he did the opening of the previous example, though interpreting it 

(595) Fr. Couperin, X Ordre, 'Fanfare', // Llvre de Pieces, Paris 
[1717], note: 

'Although the values of the treble do not seem to fit with those of the 

bass, it is customary to write thus.' 

Ex. 210. Couperin, loc. c/Y., proportional notation: 


(e) The following, however, is rightly given by Dolmetsch (loc. at.) 
as a case of conventional misnotation. 

Ex. 211. Handel, Recorder Sonata No. 4, Larghetto, (a) original 
notation, (b) interpretation (mainly after Dolmetsch, op. tit., p. 68): 

(.bar '5} 
' ' 


H-n v 
f r r i 


[liar 5) 


"1 t >' 

f ffr P 

u-i Vl l 1 *y 

- U i 


Handel has here used the two contrary methods of loose notation 
within the space of a few bars: dotted rhythm condensed to triplet 
rhythm (bar 1, etc., bass part); and equal rhythm expanded to triplet 
rhythm (bars 1 and 6, treble part, with the dot and quaver; bar 5, 
treble part, with the rest and quaver; bar 6, bass part, with the two 

Handel's habitual casualness in details of notation (see Ex. 203a in 
Ch. XLIV, 2 (a), above) prevents us from concluding that he meant here 
literally what he wrote: dotted rhythm on the first two beats of bar 1, 
'two against three' on the last beat and again on the first beat of bar 6, 
etc.; and indeed the musical result of all this confusion of rhythms is 
very unsatisfactory. Thus we have here an illustration of Quantz' tacit 



assumption at (594) above that whatever his own preference in the 
matter might be, the standard meaning of a dotted note in relation with 
triplets was triplet rhythm. It is the triplet on the last beat of bar I 
which tells the player to start in triplet rhythm. But the movement is 
also emphatically one of those which bear out C. P. E. Bach's conten- 
tion that they would be better notated in compound triple time in this 
case |. 

(f) The next two cases are best taken together. 

Ex. 212. J. S. Bach, Chorale Prelude for organ, In dulci jubilo* 
(B.G. 25, ii), misleading crochets expanded to triplet rhythm: 

Bar, 3 


Ex. 213. J. S. Bach, Organ Sonata IV (B.G. 15, pp. 46/.) semi- 
quavers expanded and dotted notes condensed to triplet rhythm; 

<a) liar 1 3 as \s ritten ( b) n iir j 5 as performed, 

A A 

(c) Bars !>Q-91 as written 


(d ) Bars 90-9 1 as performed 




In Ex. 212, triplet rhythm is notated in the alto at bar 25, but not at 
bar 3 and bar 11, etc. In Ex. 213 (a) and (c) an exactly parallel mixture 
of different notations occurs. Even viewed as a phrase in itself, the 
middle part of bar 15 is eminently eligible for expressive inequality, i.e. 
for triplet rhythm; and the fact that the top melody is in triplets con- 
firms that the performer would by now be taking triplet rhythm in both 
parts alike. But presently he runs into dotted semiquavers against the 



same triplet melody. Does he here have second thoughts, and wonder 
whether he is meant to dot, or rather 'over-dot* this new passage? 
Not so: he just goes quietly on as before, condensing the dotted rhythm 
to triplet rhythm as he was previously expanding the equal rhythm to 
triplet rhythm. 


When a movement has dotted rhythm in the notation but triplet 
rhythm in the interpretation, this includes any preliminary note. 

Ex. 214. J. S. Bach, Partita I, Courante, preliminary note in triplet 

as writttn nu\ he phm-d 

- ' > 


F. W. Marpurg (Anleitung, 2nd ed., Berlin, 1765, p. 24) demanded 
a list of composers who meant the triplets to be taken as genuine 
'two against three' and composers who did not. The first movement of 
Beethoven's 'Moonlight' Sonata is exceptional in following Quantz as 
at (594) above: Czerny makes this plain in his recollections of Beet- 
hoven's playing (see my BibL). But even as late as Liszt the standard 
interpretation as shown by C. P. E. Bach at (593) is usually better and 
often necessary in such passages. 

(596) Daniel Gottlob Turk, Klavierschule, ed. of 1802, p. 96: 
*[Two against three] is a beauty to which one has to grow accustomed/ 

The following is graphic, but its author adds the cryptic comment: 
'this elementary procedure only concerns young pupils'. 

Ex. 215. Louis Adam, Methode de Piano, Paris [1804] p. 91 : 

A 3 


New research by Michael Bruce Collins has made it desirable to 
insist still more strongly on the disappearance from baroque music of 
simultaneous cross-proportions between binary and ternary rhythms. 

In an unpublished dissertation (see my BibL), Collins has brought 



together a quantity of evidence on this most important question. He 
covers a period from 1450 to 1750. He believes that simultaneous 
combinations of ternary with binary rhythms were never practiced 
during those three centuries. 

Thus, for example, the famous mensuration canon 4 Ex una voce tres* 
by Josquin des Prez ("Agnus Dei" from Afissa UHomme arm) shows 
in its notation a tricky combination of twos against threes and fours 
against threes. It is hard to sing so, but not at all impossible, and there 
is a strange tension and beauty in the effect. But if Collins is right, the 
triplets must be squared out to duple rhythm by lengthening each first 
note, and shortening each second and third note. He may not be right. 
His argument about renaissance music is not conclusive. But his 
argument about baroque music seems to be so. 

Triplets in baroque music can be performed as such when they do 
not conflict with simultaneous duple rhythm. This is frequent. They 
can also be performed as such when the conflict can be resolved by 
assimilating the duple to the triple rhythm, as shown in Sect. 1 above. 

But the conflict may also be resolved, and sometimes must be 
resolved, by assimilating the triple rhythm to the duple rhythm. Thus 
triplet quavers may then be squared out either to quaver, semiquaver, 
semiquaver (especially French pre-1700; thereafter everywhere alike), 
or to semiquaver, semiquaver, quaver (especially Italian pre-1700; 
thereafter everywhere alike). 

Thus, for example (mine), the Tempo di Gavotta* in J. S. Bach's 
E Minor Partita for harpsichord shows occasionally in its notation the 
rhythm of semiquaver, semiquaver, quaver (to be played as written) 
but also, almost continuously, apparent triplets (every one of them to 
be played as quaver, semiquaver, semiquaver). This solution is drastic, 
but musically effective; there is no other which works through the 
movement as a whole. 

In short: throughout baroque music proper, ternary rhythm must be 
assimilated to binary, or binary to ternary, wherever the notation shows 
them apparently in simultaneous combination. 







Early music is very commonly under-phrased. This is not usually 
because the performer does not understand the phrasing, but because 
he does not realise how extremely articulate much early phrasing needs 
to be, especially in baroque music, in order to make the sense and 
structure really audible. 

Phrases are moulded by various means, which may include a dynamic 
rise and fall, a suggestion of raUentando, etc. They are separated by a 
silence of phrasing which ranges from scarcely perceptible to very 
conspicuous. The silence may be taken out of the time of the note 
before, or if the separation in the music is sufficient to justify it, added 
to the time in the shape of a momentary pause. 

A sense of phrasing is so intimate and incommunicable a part of 
interpretative musicianship that very little attempt is made to suggest 
it in notation. There are, however, a few exceptions. 


(a) The grouping of notes in ligatures partly indicated the phrasing 
under the proportional system of notation (see App. I for the more 
important effect of ligatures on note-values). 

(b) Our 'corona' sign ^ when placed over a note means that it is 
to be held longer than strict time allows; the amount varies from very 
little to very much, and is for the performer to decide. 

This sign when placed over a rest has the same effect, and may 
therefore be used to show an interval of silence added to the time. 

It may also be used over a bar-line, particularly a double bar-line, 
with the same effect. Over the final bar-line it merely indicates, some- 
what superfluously as a rule, that an indefinite silence is to follow the 
music; it may, however, serve the purpose of preventing too immediate 



an approach to an ensuing movement, in which case it is in effect the 
opposite of attacca. 

These uses are all found in baroque music, but there may be a further 
implication that a note or pause thus marked should carry free orna- 
mentation, perhaps amounting to a cadenza: for which see Ch. XII. 

Other signs were occasionally used in place of ^ and some intelli- 
gence may be needed not to misinterpret them. The following, for 
example, shows an early baroque pause sign which became a repeat 
sign later in the baroque period. 

(597) Emilio de' Cavalieri, Rappresentatione di Anima e di Corpo, 
Rome, 1600, Avvertimenti : 

'The sign *. indicates the incoronata, which is to serve for taking 
breath and to give time for making some gestures.' 

(c) Couperin used a phrasing comma, which can easily be confused 
with the main contemporary French sign (he used it himself) for trill. 

(598) Fr. Couperin, Troisteme Livre de Pieces, Paris, 1722, preface: 

'You will find a new sign of which the form is ? ; this is to mark the ends 
of melodies (chants) or of our harmonic periods (de nos Phrases har- 
moniques), and to make it understood that it is necessary to separate the 
end of a melody (chant) before passing on to that which follows. That is 
almost imperceptible as a rule, though people of good taste will feel 
something lacking in the performance if this little silence is not observed; 
in a word, it is the difference between those who read aloud continuously 
(lisent de suite) and those who stop for full-stops and commas (aux points 
and aux virgules); these silences should make themselves felt without 
altering the time (sans alterer la mesure)/ 

Couperin uses his comma to show slight separations which might 
otherwise be missed, rather than long separations which he assumed no 
good performer could miss. Unlike the modern comma, therefore, 
which is used for separations added to the time, Couperin's comma is 
to be taken out of the note before, and should often be barely per- 
ceptible. It is frequently quite superfluous, occasionally very helpful. 


(599) Girolamo Frescobaldi, Toccatas, Rome, 1614, preface, 4: 

'On the last note of the trills, or of passages of figuration by leap or by 
step, you must pause, even if this note is a quaver or a semiquaver, or 
different [in pitch] from the following note, because such a pause prevents 
confusion between one phrase and another. 

[8:] 'Before performing double passages in semiquavers for both hands, 
you must pause on the note before, even if it is a short one.' 



(600) Thomas Mace, Mustek's Monument, London, 1676, p. 109: 

'(In proper Places) . . . make a kind of Cessation, or standing still, 
sometimes Longer, and sometimes Shorter, according to the Nature, or 
Requiring ... of the Mustek.' 

(601) Joachim Quantz, Essay , Berlin, 1752, VII, 4: 

'[Flautists should] take breath at the repeat of the subject, or at the 
start of a new thought; for the end of what goes before, and the start of 
what follows, should be well separated and distinguished one from the 

[XI, 10:] 'Thoughts which belong together must not be separated; just 
as on the contrary, those where the musical sense is finished, and a new 
thought begins, without there being a rest [in the written text], must be 
separated; and this is something which must particularly be done, when 
the last note of the previous thought and the first note of the following 
thought are at the same pitch.' 

(602) C. P. E. Bach, Essay, I, Berlin, 1753, III, 28: 

'Certain notes and rests should be prolonged beyond their written 
length, for reasons of expression. 

[II, 1762, XXIX, 20:] 'At slow or moderate tempos, breaks are often 
prolonged beyond their strict length . . . this applies to pauses, cadences 
etc., as well as to breaks.' 

(603) Daniel Gottlob Turk, Klavierschule, Engl. transl. [1804], p. 34: 

'The end of a Period is best expressed by lifting the finger gently at the 
last Note, and by touching the first Note of the following Period stronger. 
By that means there arises a small Rest, which is taken from the last 
Note of the foregoing Period.' 





In proportional notation, the grouping of notes in ligatures (the 
equivalent of a modern slur) partly shows the points at which articula- 
tion occurs (i.e. between the ligatures but see App. I for the more 
important effect of the ligatures on note-values). 


Words such as cantabile and staccato, which indicate a general degree 
of articulation for the movement or passage as a whole, are not un- 
common in the later stages of baroque music. Occasionally in the 
violinist-composers an instruction for a particular technique occurs, 
such as 'attacca alia corda', indicating an attack with the bow pressed 
on to the string, the result of which is a heavy staccato or marcato 
(e.g. Vivaldi, PV 419, PV 428). 

Words giving a general indication have naturally to be applied with 
common sense, since the details of the articulation may vary very much 
within the general pattern. 


(a) The slur occurs in baroque music with increasing frequency. It 
may be used to tie notes; to show extreme legato (especially one bow 
or breath, or one word in singing); to show separate notes grouped in 
one phrase; and in conjunction with dots or dashes, to show separate 
notes taken in one bow. 

(604) John Playford, Introduction, London, 1654, ed. of 1674, p. 36: 

'A Tye is of two uses; first, when the Time is broken or struck in the 
middle of the Note, it is usual to Tye [notes across a bar-line, where 
earlier baroque notation would have divided the note-form with the bar- 
line, or put a dot after the bar-line]. 

<r The second sort of Tye is, when two or more Notes are to be Sung to 
one Syllable, or two Notes or more to be plaid with once drawing the 
Bow on the Viol or Violin. 9 

In the mid-seventeenth century English manuscript Christ Church, 
Oxford, 732, which is the Canto part of fantasies for viols in two to 



four parts by Giovanni Coperario (John Cooper) and Orlando Gibbons, 
long slurs occur which even pass across rests and are certainly not 
bowing marks. They may possibly indicate phrase-groupings, though If 
so, very casually and inconsistently. In later baroque music, we some- 
times find (e.g. Vivaldi, PV 123, middle movement) long slurs which 
are clearly phrasing marks, not bowing marks: like the long slurs in 
Beethoven's string parts, which indicate a legato intention, although 
the bow will have to be changed inconspicuously within them. 

A slur with dots or dashes is a standard sign of the later baroque 
period for notes articulated separately in one stroke of the bow. 

(605) Marin Marais, Pieces a ... Violes, II, Paris, 1701, Avertisse- 

'[A slur with dots means] to articulate all these notes, in one bow, as if 
they were played with different bowings.' 

This gives, according to the speed, many varieties of articulation 
from a deliberate staccato to a brilliant spiccato. These were virtuoso 
resources, however, and not ordinarily appropriate in baroque orches- 
tral and chamber parts. 

Plain slurs abound, but they are seldom notated consistently enough 
to be taken literally. It is for the performer to work out a consistent 

(606) Leopold Mozart, Violinschule, Augsburg, 1756, 1, iii, 16: 

'Among the signs of music the slur has considerable importance, 
although many pay small attention to it.' 

In the more brilliant violin parts, sophisticated slurrings, as used and 
taught by Geminiani, Vivaldi and others, are appropriate. In the 
average orchestral or chamber part, most of the slurring should be 
simple and obvious, and most of the slurs quite short: e.g. on groups 
of two, three or four notes, and not as a rule against the beat. This 
simplicity in the slurring is an important element in the naturalness of 
most baroque idioms. The virtuoso idioms are a separate problem, in 
which a specialised knowledge of baroque violin technique is needed; 
the ordinary idioms should not be confused with them, since here the 
articulation, though crucial, should generally be as straightforward as 
possible. There are, of course, border-line cases. 

Slurs are used on other instruments according to their respective 
techniques. They may indicate tonguing on wind instruments, and note- 
grouping on keyboard instruments. The harpsichord music of Couperin 
is particularly well served with short slurs, both with the beat and 
across the beat. 

(b) Couperin also uses a slanting stroke or strokes joining notes 
which he wants grouped closely together. 



(c) The beams connecting groups of quavers, semiquavers, etc.^ in 
place of separate tails, while sometimes used as indications of phrasings, 
are not generally so used. They may sometimes give a useful and obvious 
hint, but they should never be assumed to have musical significance 
where this is not confirmed by the sense of the music itself. They show 
great inconsistency in different copies or different parts even of the 
same work. 

(d) The standard baroque sign for staccato is the dash, of which the 
form is either | or f (either way up) without distinction of meaning. 

When the dot is used as a staccato sign, its meaning in the main 
baroque period is the same as that of the dash. At the end of the 
baroque period, however, the dot was beginning to be used to show a 
lighter, less abrupt staccato than the dash. Quantz and Leopold Mozart 
reflect this new tendency; C. P. E. Bach, however, regards the two 
signs as identical Like slurs, the signs for staccato are seldom applied 
at all consistently, and the performer should work out his own scheme 
with or without their assistance. (See Ch. XLV, 2 (h), above for the dot 
used not for staccato but in connection with inequality.) 

(607) Joachim Quantz, Essay 9 Berlin, 1752, XVII, ii, 5: 

It should be said in passing, that if there are several figures of the same 
sort of notes in sequence, and the first is marked with a slur, they must all 
be played in the same way until another kind of notes is met with. It is 
the same with notes above which there are dashes. 

[12:] 'When the dash appears instead of the dot [under a slur] the notes 
must be strongly emphasised in one bow stroke. [When there is no slur] 
the notes with dashes must be cut short, but those with dots, merely made 
with a short bow-stroke, and held on.' 

(608) C. P. E. Bach, Essay, I, Berlin, 1753, III, 17: 

'When notes are to be separated one from the other dashes or dots are 
placed over them. . . . The latter sign has been used in the exercises so 
as to avoid a confusion between dashes [which look rather like the 
figure 1] and the numbers [showing the fingering]/ 

(609) Leopold Mozart, Violinschule, Augsburg, 1756, 1, iii, 17: 

'[Dots show] that the notes covered by the slur are not only to be taken 
in one bow-stroke but have also to be separated from each other by a 
slight pressure [of the bow] . . . 

'If however small dashes are written in place of dots, the bow is lifted 
for each note, with the result that all these notes covered by the slur are 
to be taken in one bow-stroke yet have to be completely separated from 
each other/ 



Ex. 216. Louis Adam, M^thode . . . du doigtt, Paris, 1798, (a) the 
dash and (b) the dot distinguished: 

(610) Daniel Gottlob Turk, Klavierschule, Engl. transl. [1804], p. 36: 

'Lengthened, and Round Points [i.e. dashes and dots] have properly the 
same signification; but some Composers mean a shorter touch by the 
lengthened Points, than by the round ones.' 


Modem fingerings are devised to assist velocity by being interchange- 
able in all keys; early fingerings were devised to assist phrasing and 
articulation by enforcing separations where they are musically desirable. 
They may often be used, and can always be studied, to advantage. 

Two excellent examples from Couperin are cited by Wilfrid Mellers 
(Francois Couperin, London, 1950, p. 310). The slurs shown with 
dotted lines are not present in the original, but show the phrasing 
implied by the fingering. (The upper unwritten note is fingered in 
the two ornaments shown.) See also Ch. LXIV, 5, below. 

Ex. 217. Couperin, (a) *Le Moucheron', Livre II, Ordre 6, (b) 6 La 
Passacaille*, Livre II, Ordre 8, phrasing implied by the fingering: 

(611) Fr. Couperin, LArt de toucher, Paris, 1716, ed. of 1717, p. 10: 

'It is certain that a certain melody, a certain passage, when taken in a 
certain manner [of fingering], produces to the ear of a person of taste 
(de la personne de gout), a different effect/ 


The bowings shown in baroque violin treatises sometimes imply 
silences of articulation which a modern player would normally obliter- 
ate by avoiding the consecutive down-bows. Such bowings are 
occasionally useful. 



Ex. 218. T. B., Compteat Musick-Master, London, 3rd ed. of 1722, 
Ch. Ill (but see note in my Bibliography), bowings implying 
silences of articulation (u = up ; d = down) : 

u a u d li '! d 'i d d u U i d <i u ii 4 'j d u dud <i 

Ex. 219. Leopold Mozart, Violinschule, Augsburg, 1756, IV, 9, 
bowings implying silences of articulation: 


down down down down- 

[nyn ny n n y n y n nymy n] 


(a) Most problems of articulation in baroque music remain to be 
solved without assistance either direct or indirect from the notation. 

(b) Actual slurs are a normal part of the technique of bowed instru- 
ments, and are to be used even in the music of periods at which it was 
not customary to show them at all. 

(612) Diego Ortiz, Tratado, Rome, 1553,/. 3 r: 

'When two or three crotchets occur in one example, only the first is to 
be defined and the others passed over without a fresh stroke of the bow/ 

(613) Joachim Quantz, Essay, Berlin, 1752, XVII, i, 9: 

*In case there is anyone among the ripienists whose expression does not 
yet equal that of the others, the first violin ought to take him aside and 
teach him by a separate practice the true method of performance, so that 
one musician, e.g. does not . . . make slurs where the others detach the 
notes . . / 

The equivalent of slurs on wind instruments and elsewhere should 
equally be introduced into the phrasing. 

(c) The differentiation of legato from staccato is normal in any 
music. Girolamo Diruta (I/ Transttvano, Venice, 1597) attaches great 
importance to distinguishing the smooth style proper to cantabile 
music for the organ from the detached style proper to dance music on 
the harpsichord. Late renaissance and early baroque keyboard finger- 
ings (e.g. in E. Nicolaus Ammerbach, Orgel oder Instrument Tabulatur, 
Leipzig, 1571, in Diruta, op. at., and in some of the English pieces in 
the Fitzwilliam Virginal Book and elsewhere) enforce silences of 
phrasing and articulation like those of Couperin shown in 4 above. 



The entire range from staccato to legato needs to be imaginatively used 
in giving baroque phrases their point and character. 

(614) Francois Couperin, VArt de toucher le Clavecin, Paris, 1716, 
ed. of 1717, p. 61: 

'It is necessary [on the harpsichord] to sustain a perfect smoothness.' 

(615) J. S. Bach, 'Inventionen und Sinfonien', 1723, title: 
'An honest guide ... to acquire a cantabile style of playing.' 

(616) J. S. Bach, heading to Prelude and Fugue in G major for 
Organ (E.G. 38, p. 12): 

'Fuga: Alia breve e staccato/ 

The fugue thus headed is in a style highly similar to the fugal move- 
ments in J. S. Bach's violin sonatas and other works. 

(617) Anon, Advice to the Composers and Performers of Vocal 
Mustek, translated from the Italian, London, 1727, p. 14: 

'The Binding together or Stringing the Notes firm and distinct with 
the Voice, which the Italians express by the Terms (Legare and Staccare la 
Voce] are Graces equally agreable, although contrary to each other; and 
nothing but good Judgement can direct the Singer how to use them 
properly, (that is to say) according to the Nature and Design of the 

(618) Joachim Quantz, Essay, Berlin, 1752, VII, 10: 

It is necessary to study how to detect and understand well what makes 
musical sense, and what must be joined together. It is necessary to avoid, 
with equal care, separating what belongs together, and joining what com- 
prises more than one thought and should therefore be separate. For it is 
on this that one element in true expression depends. 

[XI, 10:] *It is necessary to avoid slurring notes which ought to be 
detached and detaching notes which ought to be slurred. 

[19:] 'Every instrumentalist ought to try to perform the Cantabile as a 
good singer performs it; and a good singer on his side ought to seek to 
acquire the fire of good instrumentalists with regard to liveliness, so far 
as the voice is capable of it. 

[xii, 4:] It is the quick passages in Allegro which must above all be 
performed briskly, clearly, with liveliness, with articulation and dis- 

[18:] 'Long notes ought to be given fulness in a manner characterised 
by the increase and diminution of the volume; and the quick notes which 
follow ought to be well separated in a lively idiom. 

[20:] 'But when after quick notes there follow several slow and singing 
notes, the fire must at once be restrained, and the notes performed with 



the feeling w T hich they require, in order that the hearer shall not experience 
any tedium/ 

(619) C. P. E. Bach, Essay, I, Berlin, 1753, III, 5: 

4 In general, the liveliness of allegros is conveyed by detached notes, and 
the feeling of adagios by sustained, slurred notes . . . even when not so 
marked ... I realise however that every style of performance may occur 
at any tempo. 

[6:] 'There are many who perform stickily, as if they had glue between 
their fingers. Their touch is sluggish; they hold on to notes too long. 
Others, trying to remedy this, leave the keys too soon, as if they were 
red-hot. Both are mistaken. The mean between these extremes is best, 
Here too I speak generally, since every kind of touch has its place. 

[22:] 'Notes which are neither detached, slurred nor fully sustained are 
sounded for half their value . . . Crotchets and quavers in moderate and 
slow tempos are generally performed in this half-detached style.' 

(620) Leopold Mozart, Violinschule, Augsburg, 1756, IV, 38: 

* [Lengthy passages in semiquavers should be practiced slowly to con- 
trol their speed.] Then detach the notes with shorter strokes [at greater 

(621) Giuseppe Tartini, Letter to Signora Maddalena Lombardinl, 
Padua, 1760, transl. Dr. Burney, London, 1771, 2nd ed. 1779, reprint. 
London, 1913, p. 15: 

*. . . practice, every day, one of the allegros, of which there are three in 
Corelli's solos, which entirely move in semiquavers. The first is in D . . . 
play the notes staccato, that is, separate and detached, with a little space 
between every two : for though they are written [as at Ex. 220 (a)] they 
should be played as if there was a rest after every note [as at Ex. 220 (b)]/ 

Ex. 220. Corelli, D major violin sonata, allegro semiquavers 
implying staccato in Tartini's opinion: 

I as performed K ctc 

(622) F. W. Marpurg, Anleitung, Berlin, 1755, 2nd ed. 1765, I, vii, 
p. 29: 

'Opposed to legato as well as to staccato is the ordinary movement 
which consists in lifting the finger from the last key shortly before touch- 
ing the next note. This ordinary movement, which is always understood 
[in the absence of indications to the contrary] is never indicated.' 

(623) M.-D.-J. Engramelle, Tonotechnie, Paris, 1775, p. 18: 

'All notes in performance . . . have a certain proportion of sound and 
a certain proportion of silence which together make up the total value of 
the note. These silences at the end of each note determine its articulation.* 



Engramelie's book, which Is on organ-building, was seen and 
admired by the author next to be quoted, Dom Bedos, as he was 
finishing his own book, and the two have some passages almost exactly 
in common, including (623) above. 

(624) Bedos de Celles, VArt du Facteur d'Orgues [Paris] 1766-78, 

p. 599: 

'These silences ought to vary according to the kind of expression which 
suits the piece; in lively melodies, they are ordinarily less considerable 
than in gracious ones/ 

This does not imply that 'lively melodies' are less highly articulated 
than 'gracious ones', which would be contrary to experience and com- 
mon sense; it implies that a silence of articulation, when present at all 
in 'gracious' melodies, will need to be longer, which is often true. 

(625) J. C Heck, Art of Playing the Harpsichord, London [1770], p. 5 : 

'The Gliding [Le. legato] . . . more particularly used in an Adagio. . . . 
The Staccato ... is properly to be applied in an Allegro/ 

(626) Daniel Gottlob Turk, Klavierschule, EngL transl [1804], p. 36: 

*In Notes which are to be played staccato, the fingers must be lifted, 
when almost half the Time of their length is expired, the other half is 
filled up by Rests ... 

'In Notes which are to be played legato, the finger must remain upon 
the Key, till the Time of die length of the Note is perfectly expired, so 
that not the least Separation or Rest may be heard . . . 

If Notes are to be played in the common way; that is to say, neither 
staccato nor legato, the fingers must be lifted a little sooner than the Time 
of the length of the Note is expired.' 


In the aggregate, the above passages describe a degree of articulation, 
for ordinary passages of moderate speed, which is not legato (the 
ordinary basis today) but 'opposed to legato as well as to staccato', as 
Marpurg at (622) above calls his 'ordinary movement'. 

Circumstances vary enormously; but we have the general impression 
of a very easy flow of sound, with no abrupt silence between the notes, 
yet with a certain distinctness. We tend today either to over-articulate 
baroque music, with too forceful a staccato and too little sense of line, 
or to under-articulate it, as in the modern violinist's 'detache' which 
is not detached at all. 

The problem is high-lighted where we find, as in Handel's trio sonata 
Op. II, No. 8, 2nd movt, bar 11, parallel parts in tenths one of which 
has J. J while the other has J Y J; they must be meant to sound the 
same, but what they are meant to sound is so intermediate between a 



dot and a rest that Handel could casually notate one in one way and 
the other in the other way. It is a very characteristic and revealing piece 
of inadvertence. Comparable instances abound. 

The technique of articulation differs on different instruments. But 
there is no doubt that if we can establish, particularly in the string 
department where it is most difficult, a satisfactory counterpart to 
Marpurg's 'ordinary movement*, we have gone more than half way 
towards a vital performance of the average baroque allegro. 







Volume rises with rising emotion and relaxes with relaxing emotion. 
This happens in all music, and it is a complete misunderstanding to 
confine baroque music within a range of what has recently been called 
terrace dynamics : long stretches of loud or soft flatly sustained. 

Baroque organs and harpsichords, in their dependence on hand-stops 
for changing the registration and with it the volume, lend themselves 
to terrace dynamics in some degree, owing to the difficulty of sparing 
a hand except at distinct breaks in the music; and indeed this treatment, 
within reason, brings out one of their characteristics, namely their 
imperturbable strength, very well. Other instruments have no such 
tendency; and while the structure of baroque music itself does often 
imply rather more terraced dynamics than a highly dramatic structure 
like sonata-form, it does so only on the assumption that a terrace need 
not be altogether flat. 

Akin to terrace dynamics are the popular echo-effects, when a short 
phrase is echoed pianissimo. 

But most fluctuations of loud and soft fall into neither of these set 
patterns. They have to be tailored individually to the music. 


Words to indicate loud and soft or crescendo and decrescendo 
(diminuendo) occur in a small number of fairly early baroque compo- 
sitions. They may be written either in full, or in partial abbreviation, 
or abbreviated to their initial letters. 

Claudio Monteverdi asks for a closing diminuendo to his Combatti- 
mento (first perf. Venice, 1624) in the words: questa ultima nota va in 
arcata morendo. Domenico Mazzocchi in his Dialoghi e sonetti of 1638 
uses (No. 18) the series forte . . . piano . . . pianissimo to indicate the 



same effect; elsewhere he uses words and Initial letters to indicate loud 
and soft, and crescendi and diininuendi of short extent. The words 
loud' and "soft', and their abbreviations 4 lo/ and fc so/ appear in 
English seventeenth-century manuscripts and printed books (e.g. 
Orlando Gibbons' fantasies for viols, Christ Church, Oxford, MSS. 
732-5, early seventeenth cent.; Thomas Mace, Musick j s Monument, 
London, 1676). Matthew Locke uses the words "lowder by degrees", 
"violent', and 'soft and slow by degrees' in his music for The Tempest 
(London, 1675). Fan (F) and Doux (D) are common in Lully. 

By the beginning of the eighteenth century, the Italian words forte 
and piano, etc., with a long range of abbreviations, were internationally 
current. The abbreviations are graded fromppp tofff; butpp may mean 
piu piano (i.e. more soft) as well as pianissimo. The words crescendo 
and decrescendo (or diminuendo) and their abbreviations cresc and 
deer esc (or dim) came more gradually into use; but the words crescendo 
and decrescendo were included in Leopold Mozart's list of standard 
musical terms in 1756. 

(627) Leopold Mozart, Violinschule, Augsburg, 1756, I, iii, 27: 

'Piano: means soft, and Forte, loud or strong. 

"Mezzo: means half, and is used for moderating Forte and Piano: 
namely Mezzo Forte, half strong or loud; Mezzo Piano, half soft or quiet. 

Piu: means more. So that Piu Forte is stronger; while Piu Piano is weaker 
in sound. 

'Crescendo: means increasing, and shows us that the series of notes by 
which this word is written are to increase in sound throughout. 

'Decrescendo: on the other hand, indicates that the volume of sound is to 
decrease gradually/ 


The letters p, / etc., are, in effect, used as signs of volume, and as 
we have seen in 2 above were so used in baroque music. 

The signs -*=rdH2 f r crescendo and "^^ = = sf for decrescendo 
(diminuendo) appear early in the eighteenth century (Giovanni Antonio 
Piani, Violin Sonatas, Paris, 1712; Francesco Geminiani, Violin Sonatas, 
London, 1739; Francesco Veracini, Sonate accademiche a violino solo, 
London and Florence [1744]). In some instances the hairpin is closed at 
the end : <dl] and Cll^>. 

Where /is followed shortly afterwards by p, or/? by/, the intention 
may be one of two alternatives, the choice between which can only be 
decided by the performer in the light of the music. 

Ap subito or a/subito may be intended. 

Alternatively a diminuendo from the /to the/?, or a crescendo from 
the /> to the/ may be intended. 




This point has been well established by David Boyden ('Dynamics 
in Seventeenth- and Eighteenth-Century Music', Essays on Music, 
various authors, Cambridge, Mass., 1957, p. 185) who quotes among 
other evidence the following. 

(628) W. M. Mylius, Rudimenta Musices, Gotha, 1686, cited by 
Robert Haas, Aufftihrungspraxis, Wildpark-Potsdam, 1931, p. 143: 

* With both (p and/) it is to be noted that one does not go suddenly from 
piano to forte, but one should gradually strengthen the voice and again let 
it decrease so that at the beginning p is heard, /at the middle, and once 
again p at the close.' 

Boyden (loc. cit.) also mentions the close of Mazzocchi's eighteenth 
madrigal (1638) with its diminuendo notated by forte . . . piano . . . 
pianissimo; a similar use by [Benedict] Lechler (c. 1640) of p . . . pp 
. . . ppp . . . pppp; and other clues. Walter Kolneder (Auffuhrungs- 
praxis bei Vivaldi, Leipzig, 1955, pp. 24 ff.) cites from Vivaldi/? . . .pp, 
i.e. piu piano . . . pianissimo (PV.182, V);/. . .piu/. . .jf(PV.155, 
first movt); p . . . un poco / . . . / (PV.428, last movt); etc. As 
Boyden and Kolneder correctly state, all these stand, not for a series 
of abrupt steps, but for gradual diminuendos or crescendos. 

Kolneder (loc. cit.) then lists the following grades of dynamic volume 
all indicated in Vivaldi's music by words and their abbreviations : 

TABLE XIV: original dynamic markings in Vivaldi 

pianissimo mezzo forte 

piano molto un poco forte 

piano assai / 

mezzo p /molto 

ppybp piu/ 

P ff 

quasi p 

To this he adds the pleasing comment: welch Schreckensbild fur 
einen Terrassendynamiker! ('What horror for a "terrace-dynamics- 

A further confirmation of the use of an/ or &p to indicate crescendo 
or diminuendo is found in Leopold Mozart (Violinschule, Augsburg, 
1756, XII, 8), who writes of 'the sound then diminishing again in course 
of the melody* but in his example illustrating this sentence (Ex. 221 in 
8 below) marks not a diminuendo but merely a p. The/? here stands, 
apparently, at the beginning of the intended diminuendo and is thus 
literally equivalent to dim. 

There is nothing in the notation to tell us whether an /or a/?, etc., 
in a baroque composition stands for an immediate contrast or a 



gradation. It is essential to take the context into consideration, holding 
an open mind until it is seen which interpretation makes the best and 
most natural effect. Cases of both kinds are numerous. 

A further complication is the inaccuracy with which the signs are 
placed: sometimes obviously under the wrong note; sometimes inad- 
vertently under a different note in different parts; sometimes present 
in only some of the parts; sometimes present in one copy of the same 
work but not in another copy; etc. 

On the other hand, the presence of a series such as p . . . / . . . jf, 
etc., may be taken with some confidence as an indication of a gradual 
crescendo, etc. 


Dr. Burney was so impressed by the disciplined gradation from 
extreme pianissimo to extreme fortissimo and back again which was 
the speciality of Stamitz' famous Mannheim orchestra in the middle of 
the eighteenth century that he started a legend of miraculous birth. 

(629) Charles Burney, Present State of Music in Germany, London, 
1773, I, p. 94: 

It was here [Schwetzingen, the summer resort of Mannheim] that 
Crescendo and Dimimuendo had their birth; and the Piano, which was 
before chiefly used as an echo, with which it was generally synonymous, 
as well as the Forte, were found to be musical colours which had their 
shades, as much as red or blue in painting/ 

That this orchestra was exceptional we have no reason to doubt; its 
crescendos and diminuendos were evidently the finest of the day, and 
pushed to unusual limits; but they were not intrinsically novel. Ishak, 
son of Ibrahim, both famous musicians at the court of Harun-al-Rashid 
in the eighth century, had a similar reputation for passing gradually 
through an extreme range of volume (much Oriental music does the 
same today). See (2) in Ch. I, 2 for Bishop Ethelred's twelfth-century 
complaint that 'one while the voyce is strained, anon it is remitted, 
now it is dashed, and then againe it is inlarged with a lowder sound'. 

Whether helped or not by dynamic markings, the performer has to 
produce a satisfactory range of louds and softs; of crescendos and 
diminuendos; and of fine shadings within the broader contrasts and 


(630) Christopher Simpson, Division-Violist, London, 1659, 2nd ed. 
(Division Viol), 1667, p. 10: 

'We play Loud or Soft, according to out fancy, or the humour [mood] 
of the music.' 



(631) Thomas Mace, Mustek's Monument^ London, 1676, p. 109: 

'Play some part of the Lesson Loud, and some part Soft; which gives 
much more Grace, and Lustre to Play, than any other Grace, whatsoever. 

[p. 130:] 'Humour a Lesson, by Playing some Sentences Loud, and others 
again Soft, according as they best please your own Fancy.' 

(632) Joachim Quantz, Essay 9 Berlin, 1752, XII, 23: 

'And for repeats in general, the interchange of soft and loud gives 
much grace to the playing/ 

(633) C. P. E. Bach, Essay, I, Berlin, 1753, III, 29: 

It is impossible to describe the contexts suitable to the forte or the 
piano, since for every case covered even by the best rule there will be an 
exception. The actual effect of these shadings depends on the passage, on 
the context and on the composer, who may introduce either a forte or a 
piano at a given point for reasons equally compelling. Indeed, entire 
passages, complete with all their concords and discords, may first of all be 
marked forte, and later on, piano/ 

(a) It is important, either in rehearsal or beforehand or better still 
both, to work out a definite plan of louds and softs. If they are left 
too freely to a modern player's impulses, he tends to take them half- 

(b) The basis is the structure of the movement. For example, baroque 
allegros usually open with a straightforward exposition. Very few such 
openings are given a dynamic marking at all, whereas a piano sign at 
some point after the opening is not uncommon: evidently a loud 
opening was the normal assumption. In general, forte was the standard, 
piano (or pianissimo or fortissimo) the special effect. 

Somewhere in the middle of the movement there is likely to be a 
more introverted section. The mood is quiet and contemplative, the 
modulations free and searching. A fairly consistent piano often suits 
this portion as well as a forte suits the extraverted opening. When the 
opening mood returns, the forte can return with it. These levels are 
subject to the finer shadings, but not to substantial crescendos and 
diminuendos unless the music quite definitely suggests them. 

For repeats, the simplest dynamic plan is often the best. E.g. : first 
section loud, repeat soft; second section soft, repeat loud. Or first 
section loud, repeat soft; second section, first half soft but second half 
loud, repeat the same. Or with three sections: first section loud, repeat 
soft; second section soft, repeat soft; third section loud, repeat loud. 

Da capo arias and slow movements have often a reflective middle 
section which may be taken softer on the whole than the outer sections. 

Definite echo effects, even if not marked, should be treated as such. 
They are quite common at the ends of movements. 




(634) G. Caccini, Nuove Musiche, Florence, 1602, preface, transl. in 
Playford, Introduction, London, 1654, ed. of 1674, p. 40: 

'In Encreasing and Abating the Voyce, and in Exclamations is the 
foundation of Passion.' 

(635) Christopher Simpson, Division-Violist, London, 1659, 2nd ed. 

(Division Viol, 1667, p. 10): 

6 [Loud and soft sometimes occur] in one and the same Note,' 

(636) M. Locke, Observations upon a Late Book, London, 1672, p. 36 : 

"The Viol and Violin [excell] in lowding, softing, and continuing a Note 
or Sound.' 

(637) Francois Raguenet, Comparison between the French and 
Italian Music, Paris, 1702, transl ? J. E. Galliard, London, 1709, ed. 
O. Strank, Mus, Quart., XXXII, Hi, p. 418: 

'Sometimes we meet with a swelling to which the first notes of the 
thorough-bass jar so harshly as the ear is highly offended with it, but the 
bass, continuing to play on, returns at last to the swelling with such 
beautiful intervals that we quickly discover the composer's design, in 
the choice of these discords, was to give the hearer a more true and perfect 
relish of the ravishing notes that on a sudden restore the whole harmony. 

[p. 420:] 'Every string of the bow is of an infinite length, lingering on a 
dying sound which decays gradually till at last it absolutely expires. 

[p. 426:] 'Swellings of a prodigious length . . . 

[p. 429:] 'In their tender airs [the Italian singers] soften the voice 
insensibly and at last let it die outright/ 

(638) Scipione Maffei, 'Nuova Invenzione d'un Gravecembalo', 
Giornale dei Letter at i d" Italia, V, Venice, 1711, p. 144: 

It is common knowledge among lovers of music that one of the chief 
methods by which the expert in that art contrive the secret of bringing 
particular delight to their listeners, is the piano and forte in subject and 
answer, or the gradual diminishing of the sound little by little, and the 
sudden return to the full volume of the instrument; which recourse is 
used frequently and with wonderful effect, in the great concerts of Rome 
[i.e. in the concertos, etc. of Corelli and others].' 

(639) Roger North, Autobiography, c. 1695, ed. Jessopp, London, 
1887, sect 106: 

'Learn to fill, and soften a sound, as shades in needlework, in sensation, 
so as to be like also a gust of wind, which begins with a soft air, and fills by 
degrees to a strength as makes all bend, and then softens away again into 
a temper [temperate strength], and so vanish.' 



(640) Joachim Quantz, Essay, Berlin, 1752, X, 3: 

4 Always keep the advantage of being able to produce, at need, after the 
forte & fortissimo, and after the piano a pianissimo. 
[p. 108:] 'Increase or diminish the sound as required/ 

(641) C, P. E. Bach, Essay, II, Berlin, 1762, XXIX, 12: 

'Notes which lead up to closing cadences are performed loudly [i.e. 
presumably crescendo] whether they are so notated or not. In this way 
the leader is shown that an ornamental cadence [i.e. a cadenza] is to be 
expected, for which the accompanist will wait. This indication is parti- 
cularly necessary in allegros, since ornamental cadences are more familiar 
in adagios. 

[13 :] '"When the solo part has a long sustained note which by the con- 
ventions of good performance should begin pianissimo, increase by 
degrees to a fortissimo, and return in the same way to a pianissimo, the 
accompanist must follow with the greatest exactitude/ 

(a) Like the overall contrasts of loud and soft, the big crescendos 
and diminuendos are best worked out systematically. 

(b) Again, the basis is the structure of the music. A crescendo or a 
diminuendo will sound meretricious unless it follows a genuine rise or 
fall in the emotional tension. On the other hand, if the genuine demands 
of the music are not followed merely because unmarked in an early 
text, the performance suffers from a serious loss of temperature. 

(c) Pitch and emotion generally rise and fall together. Thus rising 
sequences tend to a crescendo, falling ones to a decrescendo. Taken on 
a flat dynamic level they are apt to be monotonous. 


The effect of swelling a note and diminishing it again mentioned in 
several of the above quotations is the messa di voce of the virtuoso 
singers; but it was imitated by instrumentalists as well. It became a 
mannerism; but in really appropriate contexts it makes a beautiful 
effect. Thus the long g" with which the first violin opens J. S. Bach's 
C major trio sonata can build up to a forte, and drop to mezzo forte 
(piano would be too soft for the ensuing short notes); the second violin 
entry on c 11 can do the same; the accompaniment not being drawn into 
the messa di voce, but remaining mezzo forte. 


(642) Jean Rousseau, Traiti de la Viole, Paris, 1687, p. 23: 

The tenderness of his playing [Hauttnan on the gamba] came from 
those beautiful bowings which he brought to life and softened so skilfully 
and appropriately that he charmed all who beard him. 



[p. 56:] 'The playing of melodic pieces should be simple, and thus 
requires much delicacy and tenderness, and it is in this style of playing 
that we should especially imitate all the expressive and delightful effects 
of which the voice is capable.' 

(643) Roger North, Autobiography, c. 1695, ed. Jessopp, London, 
1887, sect. 44 ff (musical reminiscences of the later 17th cent.): 

*I was very much assisted [as a performer] by my knowledge of and 
acquaintance with the air [i.e. the art of composition]. It gave me courage 
as well as skill to fill and swell where the harmony required an emphasis.* 

(644) Joachim Quantz, Essay, Berlin, 1752, XI, 14: 

'Good expression ought nevertheless to be diversified. Light and shade 
must continually be kept up. For in truth you will never be touching, if 
you render all the notes at the same strength or the same weakness; if you 
perform, so to speak, always in the same colour, and do not know how to 
bring out and hold back the sound at the right time. Thus it is necessary to 
introduce a continual interchange of loud and soft/ 

(645) C. P. E, Bach, Essay, I, Berlin, 1753, III, 29: 

s lt is broadly true that discords are performed loud and concords soft, 
since the former arouse our feelings and the latter quiet them. A special 
turn of thought designed to arouse a violent emotion must be performed 
loudly. The so-called deceptive [cadences], therefore, whose purpose is 
such, are usually performed loudly. A rule to be noticed, wnich is not 
without foundation, is that all those notes in a melody which are foreign 
to the key can well be emphasised regardless of whether they are 
consonant or dissonant, while the notes which are indigenous to the key 
are inclined to be performed softly, regardless of whether they are 
consonant or dissonant.' [But see also (633) in 5 above]. 

(646) Leopold Mozart, Violinschule, Augsburg, 1756, XII, 8: 

Indeed, you must know how to change from weak to strong of your 
own accord, each at its right time; for this, in the familiar language of 
painters, means light and shade. The notes raised by a sharp or a natural are 
always to be taken rather more loudly, the sound then diminishing again 
in course of the melody [as at Ex. 221 (a)]. In the same way a sudden 
lowering of a note by a flat or a natural should be brought out by a loud 
[as at Ex. 221 (b)]. J 

Ex. 221. Leopold Mozart, loc. cit., fine dynamic shadings (the 
second^ in each Ex. indicating dim): 

f P 

(647) Carl Fr. Cramer, Magazin der Musik, Hamburg, 1783, p. 1217: 

*A11 who have heard [C. P. E.] Bach playing the clavichord must have 
been struck by the continual refinement of shadow and light which he 
throws over his performance.' 

(a) Unlike the overall contrasts and the big crescendos and dimin- 
uendos, the finer shadings hardly need to be premeditated. They are 
largely unconscious, and this has its dangers, since it can give rise to 
mannerisms such as the 'intrusive sforzando' with which a surprisingly 
large number of performers distort baroque music (and not only 
baroque music but perhaps it shows more there). They are, however, 
an essential part of expression, and should grow naturally enough from 
the subtle promptings of the music. 

(b) Several of the above quotations stress the importance of following 
the harmony intelligently in the dynamic shading. Quantz (Essay, 
Berlin, 1752, XVIII, vi, 12-17) elaborated a scheme for adjusting the 
volume of every chord by its degree of dissonance, with a long example 
showing markings from pp to ff on successive harmonies. Fortunately, 
he adds (op. cit., 14) that 'good judgement and sensitiveness of soul 
must also play a large part'. They must indeed: taken literally, his 
example is absurd, though as an indication of his principle it is inter- 
esting enough. A year later, C. P. E. Bach (Essay, Berlin, 1753, see 
(633) in 5, above) is almost certainly voicing his low opinion of Quantz' 
scheme when he writes that 'it is impossible to describe the contexts 
suitable to the forte or the piano, since for every case covered even by 
the best rule there will be an exception'. But, he adds, 'it is broadly 
true that discords are performed loud and concords soft'. 

Stated thus generally, this principle has almost the force of natural 
law. A discord should be perceptibly louder than its preparation, sub- 
stantially louder than its resolution. If this is overlooked, as it commonly 
is by modern performers, the progression is deprived of its natural rise 
and fall, requires more effort to follow, and sounds unnecessarily 
weighty. Even such fine dynamic shadings as this have something in 
the structure of the music for their origin. 





Balance is mainly a dynamic problem, though not entirely, since a 
part can be made to stand out by a certain significance in the manner 
of performing it, as well as by an actual predominance in volume. 


Indications of balance are occasionally found in baroque music. 
For example, in J. S. Bach's Chorale Prelude 'Christ lag in Todes- 
banden' (B.G. XL, p. 52), the melody is marked 'forte" and the counter 
subject 'piano'; his 4 Jesu giitig', Var. X (B.G. XL, p. 133) has 'Choral 
forte" ; his 'Liebster Jesu, wir sind hier', 2nd version (B.G. XXV, ii, p. 
50) has 'forte' over the melody (in two-part canon) and 'piano' over the 
lower two staves; his 'An Wasserfliissen Babylon' (B.G. XXV, ii, p. 
157) has 'piano' over the top stave, 'forte' over the chorale melody in 
the middle stave, and 'Pedale' over the bottom stave. 

Other examples are found in orchestral scores. For example, Vivaldi 
(PV. 391, middle movement) has /and 'forte sempre' marked against 
a bass line of strong rhythmic outline, &ndpo (piano) against the chords 
which accompany it on the remaining strings; and again (PV. 239, 
first movement) p followed by f (i.e. p, cresc,f)in the solo violin and 
pianissimo against the ripieno first and second violins which accompany 


Normally it is left to the performer to bring out the most important 
matter in proportion to its significance. For this to be done without 
forcing, it is equally necessary to hold back the less important matter. 

In polyphonic music, when the entries follow closely one upon 
another, it may be necessary to reduce the volume and significance of 
the part almost as soon as an entry has been made, in order to get out 
of the way of the next entry. This method is very effective in clarifying 
the more complex motets, madrigals and fantasies for viols. 

In trio sonatas, the melodic parts (not excluding the bass) should 
sometimes be equal or almost equal, sometimes take precedence in 



Solo melodies should usually stand out from their accompaniment; 
but it is not always necessary to give them much, if any, dynamic 
prominence to secure this. 

In baroque music, the bass part should nearly always be given very 
ample strength, in view of its supporting function; a point often 
overlooked in modern performance. 

(648-9) Hermann Finck, Practica Mmica, Wittenberg, 1556, Lib. V, 
[p- ?]: 

If in the beginning of the composition an elegant fugal subject occurs, 
this must be produced with a clearer and more decisive voice . . . and the 
succeeding voices, if they begin with the same fugal subject . . . are to be 
enunciated in the same way. This is to be observed in all the voices, when 
renewed fugal entries occur, so that the coherence and arrangement of all 
the fugal entries can be heard.' 

(650) Lodovico Zacconi, Prattica di Musica, Venice, 1592, LXVI, p. 

'Entries should be emphasised a little so as to be instantly and clearly 
perceived by the hearer.' 

(651) Charles Butler, Principles of Musik, London, 1636, p. 98: 

'And in their great variety of Tones, to keep still an equal Sound [with 
one another] (except in a point [of imitation]) : that one voice drown not 

(652) Francois Raguenet, Comparison between the French and Italian 
Music, Paris, 1702, transl. ? J. E. Galliard, London, 1709, ed. O. Strunk, 
Mus. Quart. XXXII, iii, p. 422: 

'Sometimes the thorough-bass lays so fast hold of our ear that in 
listening to it we forget the subject; at other times the subject is so insinuat- 
ing that we no longer regard the bass, when all of a sudden the violins 
become so ravishing that we mind neither the bass nor the subject/ 

(653) Joachim Quantz, Essay, Berlin, 1752, XII, 23: 

'When in an Allegro the subject returns several times, it is always 
necessary in the expression to distinguish well [from it] the less important 
thoughts. Whether it is majestic or caressing, gay or bold, it can always 
be made perceptible to the ear in a distinctive manner of performance 
... as also by loud and soft. 

C [XVI, 24: In a trio sonata] with regard to soft and loud, one part must 
always adapt itself to the other, so that the increase and decrease of tone 
occurs at the same time. But if it happens that for a time one part makes a 
lower part, whose notes are only such as to fill in the harmony, it must be 
performed more softly than the other which makes the principal melody 



meanwhile, so that the less tuneful passages do not stand out unsuitably. 
If the parts imitate each other, or make passages which come together 
either in thirds or in sixths; the one and the other can play at the same 

(654) Charles Avison, Essav on Musical Expression, London, 1752, 
p. 128: 

'When the inner Parts are intended as Accompanyrnents only, great 
Care should be taken to touch them in such a Marnier, that they may 
never predominate, but be always subservient to the principal Performer, 
who also should observe the same Method, whenever his Part becomes 
an Accompanyment; which generally happens in well-wrought Fugues 
and other full Pieces, where the Subject and Air are almost equally 
distributed. When the Attention of every Performer is thus employed by 
listening to the other Parts, without which he cannot do Justice to his 
own, it is then that we may expect to hear the proper Effect of the whole.* 

(655) Jean-Jacques Rousseau, Letire sur la Muslque Franfoise [Paris], 
1753, p. 53: 

1 am well aware that the bass, being the foundation of all the harmony, 
ought to dominate the rest of it, and that when the other parts stifle it or 
cover it, the result is a confusion which can obscure the harmony; and 
that is how I explain to myself why the Italians, sparing of their right 
hand in accompanying, usually double the bass at the octave in their left; 
why they put so many double basses in their orchestras; and why they so 
often make their tenors [violas] go with the bass, instead of giving them a 
separate part, as the French never fail to do/ 

(656) [Laugier] Apologie de la Musique Franfoise [Paris] 1754, p. 69: 

'The first care which ought to be given is to arrange the consort in good 
order, to furnish all the parts sufficiently, to see that each makes its effect, 
that the chief parts, such as the treble and the bass, stand out to advantage, 
and that the subsidiary parts, such as the counter-tenor (Haute-contre) 
and the tenor (Taille) are less conspicuous, so that a harmony results in 
which nothing grows out of bounds, and which has unity. One cannot 
too strongly recommend furnishing the basses beyond the rest; for they 
are the foundation of the harmony, and from the nature of low-pitched 
sound, always less penetrating [a strong bass-line being especially neces- 
sary in recitatives].' 

(657) Charles Burney, Present State of Music in Germany -, London, 
1773, 1, p. 25: 

'[A praiseworthy opera of which] the accompaniment is both rich, 
ingenious, and transparent, if I may be allowed the expression, by which 
I mean, that the air is not suffocated, but can be distinctly heard through 



(657a) J. F. Reichhardt, Vber die Pflichten des Ripien-Viottnisten, 

Berlin and Leipzig, 1776, Ch. 11: 

'In fugues ... on long notes, the bow must be started somewhat more 
firmly, and a certain controlled stress of the notes is necessary here in 
order to keep the interpretation clear. This is the more necessary the more 
parts the fugue has. In particular the subject, at each new entry, has to be 
brought out with a certain pressure of the bow.' 





Accentuation was not shown directly in baroque notation; but instruc- 
tions are sometimes found which indicate a well-accented style. Such 
instructions are not uncommonly attached to movements which might 
otherwise be taken as cases for expressive inequality, as discussed in 
Ch. XLV above: e.g. 'notes egales et marquees'; 'notes martelees* 
(hammered notes); 'detachez'; 4 mouvement decide'; "mouvement 
marque*; etc. Words such as staccato, spiccato or spiritoso, all sug- 
gestive of accentuation in some degree, are likewise found, especially 
in the Italian virtuoso violinist-composers. 


It is a universal necessity to respond with the appropriate accentua- 
tion to the implication of the music. 

(658) Rene Descartes, Musicae Compendium, (1618) 1650, transl 
William Viscount Brouncker, London, 1653, Ch. Ill, p. 5: 

'Few have understood, how this Measure can be exhibited to the ears 
without a percussion, or stroke, in Musick, very diminute [florid] and of 
many voices. This we say is effected only by a certain intension of the 
Spirit or breath, in Vocall Musick; or of the Touch, in Instrumental: so as 
from the beginning of each stroke, the sound is emitted more distinctly. 
Which all Singers naturally observe, and those who play on Instruments; 
principally in Tunes, at whose numbers we are wont to dance and leap. 9 

In late renaissance and, less frequently, in early baroque music, we 
may find complete polyphonic freedom in the accentuation implied by 
the separate parts; and even in baroque music as a whole, where 
barring has become normal usage and mainly coincides with the pre- 
vailing measure, the requisite accentuation is often far more subtle and 
irregular than appearance suggests. The reader is referred to Gemini- 
ani's warning, quoted at (536) in Ch. XXXEX, 1 above, against crudely 
accenting the first beat of every bar. It is the shape of the phrase which 
implies the accentuation. Every case has to be judged on merits by 
ordinary good musicianship ; but very often the highest note of a phrase 
is also its musical climax, and where this is so, the accent usually falls 



in the same place too. A somewhat crude case of this is shown by 
Leopold Mozart below. Albert Schweitzer's discussion of the subject, 
in his J. S. Bach (Paris, 1905 and Leipzig, 1908) is worth consulting. 
(659) Leopold Mozart, Violinschuk, Augsburg, 1756, XII, 13: 

'In lively pieces the accent is generally taken on the highest notes, so 
as to make the performance really cheerful* 

Ex. 222. Leopold Mozart, loc. cit. 9 accents on high notes off the 


The term 'agogic accent' has been adopted for a slight prolongation 
of a note by which it acquires a certain prominence, whether or not 
this agogic accent is combined (as it usually is) with an actual dynamic 

The agogic accent was familiar to baroque musicians, though not 
under that name. In so far as it required description at all, it was 
included under the principle of inequality. Quantz* account of inequality 
at (587) in Ch. XLV, 7 above, after mentioning the normal convention 
by which 'the first and third notes' of four are to be 'a little pressed 
upon, and their tone made a little louder than that of the second and 
fourth notes', adds: 'but we except from this rule in the first place 
quick passages, in a very quick movement, where it is only possible to 
press and use force upon the first note of the four', etc., etc. The term 
'press upon* is here equivalent to 'prolong slightly' ; the term 'use force 
upon' is equivalent to 'accent'. Quantz is thus describing what is 
actually a combination of agogic and dynamic accents. 


On the violins, a weight accent is made literally with the weight of 
the arm as it presses or even drops the bow on the string. Other 
instruments can produce comparable effects. 

There are many gradations of weight accent; but it tends to be too 
massive for most baroque contexts. 


This is a slightly less abrupt access of weight just after the note has 
started. As a perpetual unconscious mannerism it is particularly 
harmful in baroque music, but as an intentional dynamic accent, 



particularly useful Milder than the weight accent, it gives just the right 
emphasis for the peak of a phrase. It combines well with an agogic 
accent. It can be prepared by the slightest of deliberate hesitations. It 
is capable of fine gradations. 

(660) Giulio Caccini, Nuove Muslche, Florence, 1602, preface, Eng. 
transl. in Playford, Introduction, London, 1654: 

'There are some that in the Tuning of the first Note, Tune it a Third 
tinder: Others Tune the said first Note in his proper Tune, always increas- 
ing it in Lowdness, saying that this is a good way of putting forth the 
Voyce gracefully ... I have found it a more affectuous way to Tune the 
Voyce by a contrary effect to the other, that is, to Tune die first Note, 
Diminishing it: Because Exclamation is the principal means to move the 
Affections; and Exclamation properly is no other thing, but in the 
slacking of the Voyce to reinforce It somewhat . . . 

'Exclamations may be used In all Passionate Musicks/ 

(661) Joachim Quantz, Essay, Berlin, 1752, XII, 19: 

* When after quick notes there suddenly comes a long note which holds 
up the melody, it must be marked with a special emphasis, and likewise in 
the following notes the volume can again be restrained a little/ 

(662) Leopold Mozart, Violinschule^ Augsburg, 1756, XII, 8: 

It is the custom always strongly to emphasise minims mingled with 
short notes, and to diminish the sound again [as at Ex. 223 (a)]. Indeed, 
many crotchets are performed in the same way [as at Ex. 223 (b)]/ 

Ex. 223. Leopold Mozart, loc. cit., sforzandos: 

* ^fp 3 f p fp fp fp fp 


(a) The crystalline sharpness and transparency so characteristic of 
baroque music are well served by the attack accent, especially on string 
instruments. The player uses his forefinger to press the bow momen- 
tarily on to the string; as he releases it into motion, a beautifully crisp 
and clean attack should result. Wind instruments can do the same by 

This, too, can be finely graded. At its least, it is an incisive attack, 
but no more; in this form, it can be used on every note of a passage 
which needs to be sharply etched without being heavy. Taken more 
strongly, on selected notes, it is perceived not only as an attack, but 



as a subtle accent, so that any note or notes in the passage can be 
picked out with no sense of effort and no interruption of the flow. 
Combined with a certain weight borrowed from the weight accent, but 
in moderation, the attack accent can be given sufficient power for 
almost any baroque situation. 

The effect of an attack accent can be greatly heightened by a little 
silence of articulation taken out of the time of the note before. The 
silence alone can give an impression of accentuation on the harpsichord, 
where there is scarcely any true dynamic accentuation, or on the organ, 
where there is none. The combination of preparatory silence and true 
attack accent can give the rhythm a quite intoxicating lift, and is the 
most generally valuable of all baroque methods of accentuation. 

(663) Leopold Mozart, Violinschule, Augsburg, 1756, IV, 23: 

'You must not forget to attack the middle [syncopated] note more 
strongly with the up bow-stroke; and to slur the third note to it with a 
gradual diminution of the volume/ 

Ex. 224. Leopold Mozart, op. cff., 22, silence of articulation 
shown before (accented) syncopation: 

down up clown up 








There is an Intimate connection between music and its instruments, 
The composer more or less consciously takes into account their 
sonority, the attack and decay of their sounds, their natural articulation 
and their range of technical resource. He also exerts more or less 
pressure on performers to extend their technique and on makers to 
develop or modify the instruments themselves; and performers and 
makers have their own urge to evolve in new directions, often advan- 
tageously, though not always without a loss to weigh against the gain. 

The climate of the times may vary greatly. In the sixteenth century, 
there was a disposition towards polyphonic patterns which can be 
traced in different sonorities with no intrinsic difference of effect. This 
did not make sonority a matter of indifference; musicians took an 
almost extravagant delight in experimenting with rich and adventurous 
combinations of instruments. But, within appropriate limits, we can 
choose our own sonority for modern performance, provided only that 
we make it no less colourful than it would then have been. 

A case, again, such as the Art of Fugue, the prodigious contrapuntal 
pattern on which J. S. Bach was working when he died, comes perhaps 
as near as music can come to abstract line; and it is not even known 
what instrumental sonority, if any, Bach conceived for it. Some writers 
have regarded it as music for the mind alone. But is music ever for the 
mind alone? Can mind and matter ever be divorced? The Art of Fugue 
can be made effective on the harpsichord, on a string quartet or a 
consort of viols, in a mixed scoring of strings and wind, and in other 
ways. It is only necessary to provide it with a sensitive reflection in 
sound of the thought and the pattern in the music. We are not confined 
to one sonority provided we keep our sonority appropriate. 

Contrast with this the music written by J. S. Bach specifically for the 
keyboard instruments. Of these, the organ is in a class apart: works 
really well composed for the nature of this remarkable instrument are 
unmistakable. Harpsichord music is not quite so distinct; music desig- 
nated for the clavier ('keyboard') in J. S. Bach's Germany was more or 
less interchangeable between the harpsichord and the clavichord, as 
becomes abundantly clear, for example, in C. P. E. Bach's great Essay 



of 1753. Yet the sonority of the two instruments, though similar, is 
not the same; while their attack and technique are not even very 
similar. The one excels in brilliance, the other in expressiveness. Of 
Bach's clavier music, some examples are far more effective on the 
harpsichord, other examples on the clavichord, others equally on either, 
but none are quite unsuitable on either, because the common qualities 
of the two instruments are more important musically than their 

Now consider the position of the piano with regard to baroque 
music. J. S. Bach was not favourably impressed by Silbermann's early 
pianos. They were, of course, imperfect instruments; but Bach would 
have been quick to see the possibilities behind the imperfections, if 
these possibilities had been of keen interest to him. They were not of 
keen interest, because they were quite foreign to his own keyboard 

The possibilities latent in the piano were not merely its continuous 
gradations of dynamic control (the clavichord has these, though at a 
lower absolute volume) but its impressionistic surge of sound. The 
former possibilities were developed first: the pianos of Mozart's genera- 
tion, though true pianos in dynamic control, are nearer to the harpsi- 
chord than to the modern piano in sonority, and also in clarity. The 
impressionistic surge is rather more developed in the pianos of Beet- 
hoven and Schubert; but not until Schumann and Brahms did composers, 
performers or makers exploit this aspect to its fullest extent. The 
contrast between Bach's clear counterpoint and Schumann's surging 
arpeggiation is precisely matched by the contrast between the mid- 
eighteenth century harpsichord and the mid-nineteenth century piano. 
There is no clearer instance in history of a change in the instruments 
of music evolving in mutual inter-reaction with a change in the 
prevailing aims and climate of music itself. 

Thus while the choice of harpsichord or clavichord for Bach's 
clavier music is to a large extent a matter of taste or even of convenience, 
the use of the piano raises a nice point of artistic suitability. The pro- 
found integrity, the warmth and the intelligence of Harold Samuel's 
Bach playing will still be remembered by many; Myra Hess is just as 
enjoyable now; Rosalyn Tureck, though too puritanical for my taste, 
is enjoyed by many; and it is out of the question to dismiss such artists 
as despoilers of Bach's original intentions. Yet it is obvious that they 
are departing from the original. Music composed for the even flow and 
brilliant clarity of the harpsichord has to be mentally reconceived 
(transcribed would hardly be too strong a word) when interpreted in 
terms of the dynamic flexibility and massive impetus of the piano. It is 
probably better, at any rate in solo music, to treat the piano pianistically 



and make the most of its own fine qualities, rather try to It 

sound like a harpsichord, which it can never do. To suppress its proper 
warmth Is to spoil it and denature It, and does justice neither to the 
Instrument nor to the music. 


Substitutions were by no means foreign to the baroque mentality. 
Baroque title-pages do not usually specify quite the free choice of 
alternative instruments found regularly on renaissance title-pages; but 
the option is still often very wide. Solo and trio sonatas were commonly 
published for a choice of violin, flute or oboe. There was a fashion in 
operatic excerpts to be played on various Instruments: e.g. Ballard's 
excerpts from Lully's operas 'suitable for singing and playing on the 
Flute, Violin and other instruments' (Paris, early eighteenth century). 
The same freedom extended to other vocal and instrumental music. 

(664) Michel L'Affiilard, Principes tres-faciles, Paris [71694], ed. of 
1722, p. 6: 

If people who play instruments wish to play the "Airs de mouvement" 
in this book, they have only to transpose them into the key which best 
suits the compass of their instruments.' 

(665) Frangois Couperin, Concerts Royaux, Paris, 1722, preface: 

'They suit not only the harpsichord, but also the violin, the flute, the 
oboe, the viol and the bassoon. 

'[Troisieme livre de pieces, Paris, 1722, preface:] These pieces, indeed, 
are suitable for two flutes or oboes, as well as for two violins, two viols, 
and other instruments of equal pitch; it being understood that those 
who perform them adapt them to the range of theirs.' 

(666) Jean-Philippe Rameau, Pieces de clavecin en concerts, with a 
violin or a flute and a [bass] viol or a second violin, Paris, 1741, preface: 

'[When substituting the flute] if one finds chords, it is necessary to 
choose the note which makes the most beautiful melody, and which is 
ordinarily the highest. 

'With regard to notes which pass beyond the compass at the bottom of 
the flute [take the passage an octave up, between the signs shown; but] in 
a rapid passage of several notes, it is sufficient to substitute for those 
which descend too low the neighbouring ones in the same harmony, or 
to repeat those which one deems fit.' 

The motive behind such titles and instructions was only partly the 
obvious commercial one of increasing sales. There was also a genuine 
preference for leaving the choice open, for providing music in such a 
form that as many musicians as possible could use it, and for trusting 



them to do so with Intelligent discrimination. If they made an undis- 
criminating choice, the loss was theirs. That remains true. Some pieces 
are extremely idiomatic, others are more accommodating. Some are for 
named instruments, and obviously require them. Where alternatives 
are stated, they do not necessarily all apply to the same piece. Often 
only one of them is really suitable. Conversely, where no alternatives 
are stated, this does not exclude them, provided that the same care is 
taken to suit the choice of instruments to the music in question. 

There is no fundamental reason why the alternatives should be 
limited to the strict historical possibilities; though as most of the 
suitable ones already existed, this will ordinarily happen. The only 
fundamental issue is the issue of suitability. If a saxophonist finds a 
baroque sonata which makes an excellent musical effect on his instru- 
ment, which was invented in the nineteenth century, there is nothing 
in the baroque attitude to music which prevents him. On the other hand, 
if a small heckelphone (invented in the twentieth century) or a clarinet 
is used to replace a long D trumpet in a high and difficult baroque 
obbligato part, the musical effect is not adequately served : partly because 
it is in this context inferior as well as different; partly because we have 
here an instance not of optional timbres but of specific orchestration, 
so that the mere fact of sounding different is detrimental. If Bach's 
sixth Brandenburg concerto is taken with additional violas or cellos in 
place of gambas, its already dark colouring becomes disadvantageous^ 
sombre, and this imaginative experiment in chamber-music scoring (it 
is not really an orchestral work) suffers considerable loss of beauty. 
Yet such substitutions cannot always be avoided under prevailing 
conditions, and it is possible to minimise the disadvantages. 


We have also to consider the very substantial changes of technique 
found in many instruments whose basic form has not changed, and 
which are still in everyday use. These and other points will be briefly 
considered in the following chapters, though for an adequate treatment 
a very much fuller study would be required than I am able to give them. 





The pitch at which music is performed makes a great difference to its 

Too high a pitch tends to a strident effect, too low a pitch to a 
sombre effect. Singers, indeed, can be pushed not only out of their best 
tessitura but out of their compass by a relatively small change in pitch. 
Instruments may be less obviously affected : but consider the difference 
in colouring between a B flat clarinet and an E flat clarinet! 


(a) Here we have to be careful not to confuse actual pitch with 
nominal pitch. 

Our present international standard of pitch is a 1 = 440. This means 
that we assume for our nominal pitch a 1 an actual pitch given by 440 
c.p.s. (double vibration-cycles per second). 

No such assumption can be made for earlier periods, which related 
an astonishing variety of actual pitches to the nominal pitch a 1 . We 
have abundant evidence on the nominal pitches of early instruments, 
etc., but very little on the actual pitches, and much of what we have 
depends on calculations liable to error. 

(b) We have also to allow for the effects of transposition, which was 
a common recourse to meet the numerous conflicting actual pitches in 
use. We may, for example, know the actual original pitch of a certain 
organ, at least approximately; yet the music performed with it may 
still have been performed at a quite different pitch, the organist trans- 
posing his part at sight. The same applies to woodwind or brass 

When the three clefs (G, C, and F) are used on higher or lower 
positions of the stave than their usual S.A.T.B. positions, then in vocal 
polyphony of the late renaissance and early baroque periods, trans- 
position may be implied the commonest being a fourth or a fifth 
downwards (where the clefs stand high). Some authorities describe 
this as normal practice; others warn us against spoiling an intentionally 
low and sombre tessitura, or an intentionally high and brilliant tessitura, 
by such transposition. We are therefore thrown back on our musical 



judgment of what the intended tessitura really is in any given case, 
except where a composer's preface specifies a transposition or a trans- 
posed organ part exists in writing; and even there, we are still dealing 
in nominal pitches, without necessarily knowing to what actual pitch 
they correspond. The problem (known since the later eighteenth century 
as that of the chiavetti, little clefs, or chiavi trasportati, transposed 
clefs, as opposed to chiavi naturali^ natural clefs) has been studied 
with great thoroughness and insight by Arthur Mendel ('Pitch in the 
16th and early 17th Centuries', Mus. Quart., XXXIV, 1948, Nos. 3 and 
4). His articles are essential reading for any musician directing vocal 
polyphony of the renaissance and early baroque periods. But they 
confirm only too clearly what has just been said: we must be prepared 
for extensive transposition, yet have usually to rely on purely musical 
considerations in deciding where, and by how wide an interval, to 
apply it. 

Clefs also appear at other than their usual S.A.T.B. positions on 
the stave for convenience in avoiding ledger lines (to which early 
musicians cherished an exaggerated aversion) without any possible 
implication of transposition. 

The G clef on the lowest (instead of the usual lowest but one) line 
of the treble stave (Trench violin clef) was associated with French 
notation, and its presence is sometimes an indication that the music is 
in the French style (see Ch. II, 3, above). 

(c) A remarkably painstaking attempt to clarify the situation with 
regard to early pitch was made by A. J. Ellis, culminating in a historic 
article ('History of Musical Pitch", Journal of the [Royal] Society of 
Arts, London, 5 March, 1880), which has been the foundation of all 
subsequent attempts. Ellis gave his results to one decimal place, in 
c.p.s. ; and though he carefully explained the circumstances which made 
most of them subject to a margin of error, many writers who have since 
drawn on his article have ignored the explanation and copied the results, 
so that a mistaken assumption has become current that the pitch of 
some early organs, for example, can be stated with this degree of 
accuracy. Ellis himself worked largely on the pitch of actual organ 
pipes (which may have been changed at an intervening date) and on 
recorded measurements (but calculation of pitch from measurements is 
complicated by the need to make various corrections which cannot be 
exactly estimated, and of which even the theory was in one or two 
respects unknown to Ellis). 

Not till the development of the tuning fork, early in the eighteenth 
century, did evidence of actual pitch appear which is both exact and 

(d) The important series of articles by Mendel cited above, and 'On 



the Pitches In Use in L S. Bach's Time', (Mus. Quart., XLI, 3 5 4) 
greatly added to our understanding of the problem. Two crucial 
were clarified by Mendel's researches: the immense variety of nominal 
pitches of the late renaissance and baroque periods; and the narrow 
range of actual pitches after transposition has been taken into 

(e) A brilliantly lucid account, both scientifically and musically 
reliable, was contributed by LI S. Lloyd to Grove's Dictionary (5th 
ed., London, 1954, s.v. 'Pitch, Standard"). This is now much the best 
introduction to the subject for the general reader. 



(a) Michael Praetorius (Syntagma, II, Wolfenbiittel, 1619) records 
dimensions of a pipe in the Halberstadt organ, which he examined in 
the state in which it had been left when reconstructed in 1495. Ellis 
made a calculation which, taking into account the criticisms of his 
methods by Mendel and Lloyd, we may translate as meaning that the 
pitch of this organ stood at about a 1 = 550 to about 630. This is four 
to six semitones above our present a 1 = 440. 

By a further calculation, based on information in Arnolt Schlick's 
Spiegel der Orgelmacher, [Speyer, 1511], but seriously questioned by 
Mendel, Ellis inferred an almost identical pitch which he took to be 
Schlick's high pitch, and another nearly three semitones below a 1 = 
440, which he took to be Schlick's low pitch (particularly doubtful). 
Ellis further gives an organ at St. Catherine's, Hamburg, as a semitone 
and a half above a 1 = 440. 

(b) Canon E. H. Fellowes (English Madrigal Composers, Oxford, 
1921, pp. 70-74) argues the question of pitch in late Elizabethan and 
early Stuart England in a very practical and valuable manner. His chief 
evidence is the written compass and tessitura of voice parts; he also 
cites a little acoustic evidence (particularly Thomas Tomkins' mention 
of a 30-inch organ pipe at the nominal pitch of F, which might sound 
between G and A flat at a 1 = 440). 

Fellowes concludes that Tudor church music was composed for an 
actual pitch just in excess of a tone above a { = 440, and should be 
transposed a tone up in performance to obtain the effect actually 
intended by the notation. 

He concludes that secular music of the late sixteenth and early 
seventeenth centuries in England stood at a nominal pitch about 
equivalent to a 1 = 440, so that it requires no transposition, the actual 
standard of pitch being the same as ours. 



He points out that these pitches would have been subject to con- 
siderable variation. There seems no reason to doubt the general 
correctness of Ms views. 

He believes that there was another pitch a minor third below a 1 = 
440, for virginals, of which the surviving specimens would not in his 
opinion stand being tuned higher. Here he was on much less secure 
ground: so much depends on condition, string-gauges and other 

(c) The instructions in tutors for the lute, the viol, etc., are to tune 
up the gut top string to just short of breaking point. To find the point 
at which the string breaks in your face is in no way difficult; the point 
just short of it is, shall we say, less painfully obvious. But with good 
strings of the same gauge relative to their length it is more constant 
than might be thought, Michael Prynne (Grove's Dictionary, 5th ed., 
London, 1954, s.v. 'Lute') after experiments with the lute reached the 
tentative opinion that this instrument may ordinarily have been tuned 
no higher than three-quarters of a tone below a 1 = 440. But there are, 
again, too many variables for these experiments to be conclusive. 

(d) Anthony Baines (Woodwind Instruments and their History, 
London, 1957, p. 242) in course of the most valuable study yet written 
of old and new woodwind, suggests from the examination of surviving 
instruments that some approach to a standard pitch of about a 1 = 470 
or roughly a semitone above a 1 = 440 can be seen among the leading 
woodwind makers of the sixteenth century. 

(e) Pitches given by Ellis for the first half of the seventeenth century 
include one for 'Mersenne's spinet, 1648', well over a semitone below 
a> as 440; an organ pitch recommended by Praetorius (1619), about 
half a semitone below a 1 = 440; another from the Great Franciscan 
Organ, Vienna (1640), nearly a semitone above a [ = 440; two from 
Mersenne (1636), a ton de chapelk (or Chorton: church or choral pitch) 
well over a tone above a [ = 440, and a ton de chambre (or Kammerton; 
chamber pitch) well over two tones above a [ = 440. 


(a) Ellis gives 'Eulefs clavichord, 1739*, a tone below a 1 = 440; a 
Silbermann organ at Dresden, 1754, a semitone below a 1 = 440; a 
Father Smith organ, 1690, at approximately a 1 = 440, and another, 
1683, at well over a semitone above a 1 = 440; an organ at Rendsburg, 
1700, just over a tone above a 1 = 440; etc. Varieties of Chorton and 
Kammerton high, and low still abounded. 

Lloyd states that many English organs were lowered in pitch, some- 
times by as much as a tone, between 1650 and 1750, with the result, 
he says, that church pitch came down to secular pitch. This historical 



evidence of work actually done on organs is solid confirmation for 
Fellowes* assumption, mentioned In 3 (b) above, that church was 
higher than secular pitch In Elizabethan and early Stuart England. 

(b) By the end of the baroque period we begin to have the 
evidence of tuning-forks (the pitch-pipe previously in use is an Inexact 

A tuning-fork said to be Handel's was examined by Ellis ("Handel's 
pitch, 1751') and found to give a pitch rather less than a semitone below 
#J = 440; and Lloyd concludes that one locally used pitch of Handel's 
time might be this or a little lower: In broad terms, let us say, a semitone 
below a 1 = 440. We have no evidence that the fork was Handel's. 

(c) Anthony Baines (loc. cit.) believes from the examination of very 
large numbers of specimens that woodwind makers of the dominant 
French school of Hotteterre and others during the second half of the 
seventeenth century arrived at a moderately standard pitch of about 
a\ = 422 or roughly a semitone below a 1 = 440, and that the Influence 
of their famous instruments kept orchestral and chamber pitch reason- 
ably constant for about a century or rather longer, at this low pitch' 
a semitone down on ours. 

(d) Yet English recorders by Bressan are in existence, which were 
made between 1700 and 1730, and of which the pitch is in most cases 
nearly a tone below a\ = 440. 


(a) A copy of the tuning-fork to which Mozart's piano may have 
been tuned was seen by Ellis ('Mozart's pitch, 1780') and found to give 
rather less than a semitone below a 1 = 440. There is no evidence that 
this constituted a 'Viennese Classical Pitch'. 

(b) Ellis further gives: 'Original Philharmonic Pitch, 1813', at just 
over half a semitone below a { = 440; 'Sir George Smart's fork, 1820- 
26', at rather less than a third of a semitone below a 1 = 440; 'Scheibler's 
Stuttgart Standard, 1834' at a* = 440; 'Madrid, 1858', 'San Carlo, 
Naples, 1857', 'Broadwood's Medium, 1849' and Trench opera, 1856', 
all at just over a 1 = 440; 'Highest Philharmonic, 1874*, 'Broadwood, 
firard and Steinway, England, 1879' and 'Vienna, high, 1859', all at 
just over half a semitone above a 1 = 440; 'St. James, Hamburg, 1879', 
at a tone above a 1 = 440. 

(c) 'Scheibler's Stuttgart Standard, 1834' refers to a most interesting 
attempt by this notable pioneer to secure a standard pitch, which he 
proposed to a congress of physicists at Stuttgart in 1834. His suggestion 
of a 1 = 440, with its striking anticipation of our present level, was 
arrived at as the mean pitch of Viennese grand pianos taken at a 
variety of temperatures (because of the effect of rises of temperature in 



raising the pitch of most musical instruments), as they were tuned at 
that date. 

(d) In 1858 the French Government appointed a commission of 
distinguished musicians, physicists and civil servants to consider the 
chaotic situation in which pitch still stood. Its report led in 1859 to two 
decrees giving legal force to a standard pitch of a 1 = 435, just below 
a\ = 440. This is diapason normal. 

This pitch was accepted to some extent in other Continental countries. 
In England, it was misunderstood, and the actual task of preparing a 
standard tuning-fork was incompetently handled; hence the Society of 
Arts, which had meant to establish a 1 = 440, found itself with a C fork 
giving 534-5 c.p.s., which produces (in equal temperament) an a 1 of 
449-5 c.p.s. about a quarter of a semitone above a 1 = 440. 

The perennial fantasy of improving the brilliance by raising the pitch 
must have been particularly active in England at this date; for in May 
1877, Wagner was enraged by the strain his singers were put to by 
being obliged to sing in London at a pitch much higher than usual, 
while in 1879 Patti actually refused to sing at Covent Garden at the 
pitch its orchestra had then reached, which was a 1 = 455, more than 
half a semitone above a 1 = 440. The orchestra came down to French 
pitch. The power of a great singer is not always misused. 

(e) In 1885, an international conference on pitch was held at Vienna, 
at which, however, England was not represented. It accepted, in effect, 
d = 435. 

(f) In 1895, Robert Newman, the manager of Queen's Hall, and his 
young conductor Henry Wood, had the courage, regardless of the 
expense, to lower the pitch of the newly built organ. The Philharmonic 
Society responded with the admirable if belated decision to adapt the 
French pitch, the determinants of which they nevertheless still suffi- 
ciently misunderstood (by relating it to temperature, which had been 
deliberately avoided in France) to arrive, not at a 1 = 435, but at 
a 1 = 439. This is New Philharmonic Pitch: virtually a 1 = 440. 

Any upward lead in pitch is all too readily followed, and the English 
mistake was unfortunate. 

(g) In 1939, a fully representative international conference in 
London, organised, it is pleasant to record, by the British Standards 
Institution, adopted our present a 1 = 440, which had become standard- 
ised in America, and is a much more convenient frequency than a ] = 
439 for calculating mathematically and producing electronically. This 
is British Standard Pitch, and is generally accepted, though attempts to 
edge upwards again are reported. It is important that this pitch should 
be held, since there are real musical disadvantages in a higher one, and 
it is in any case a little on the high side. 



Much the most exact and dependable standard by which to tune is 
an electronically produced tone such as that broadcast dally by the 
B.B.C. at a 1 = 440. A tuning-fork is sufficiently reliable in practice; an 
oboe A is not. 


(a) The evidence as re-examined by Lloyd, and much more exhaus- 
tively by Mendel, does not permit us to believe that the pitch of about 
a 1 = 422, roughly a semitone below a 1 = 440, and nowadays known as 
'low pitch', had any greater currency than other pitches during the 
baroque period. It was, indeed, a pitch in use, but only as one among a 
wide variety of others, and we have no grounds for giving it any general 
preference. It is sometimes a useful pitch, but no more than that. 

(b) Any general statement that viols or harpsichords, etc., were built 
to be tuned at 'low pitch' roughly a semitone below a 1 = 440 is thus 
against the evidence. In relation to the bulk of the repertoire, both 
classical and pre-classical, our present a 1 = 440 is a little too high to 
be ideal, but it is not far out, and the advantages of having at last 
secured an agreed standard are far too valuable to forego by adopting 
another pitch or pitches for early music. Under all ordinary circum- 
stances, therefore, and except for special reasons, we should keep to 
0i =440. 

(c) We must, however, make use of transposition whenever this is 
necessary to bring the music within the compass of the human voice, 
or within the best tessitura having regard to the character of the music 
so far as we can judge it. 

We may also make use of transposition if we suspect, even apart 
from vocal considerations, that the nominal pitch at which we find the 
music differs substantially from the actual pitch which its notation 
gives at a 1 = 440. This suspicion may often amount to no more than 
a conviction that the music sounds more its natural self when trans- 
posed, say, a fourth up, and we have to be careful then that we are 
not merely superimposing our own predilections at the expense of the 
composer's. But not too careful. His music is almost certain to have 
been performed at different pitches by different performers of his own 

(d) Except where required to meet the needs of singers, small trans- 
positions are sometimes to be avoided because of an adverse, or at 
least a transforming, effect on the sonority. Violins, viols or lutes lose 
some of their brightness when transposed, e.g., from D major to E flat 
major, which has fewer notes enriched by sympathetic resonance from 
the open strings. On the other hand, transposition from F major to 



G major makes less difference in this respect, and can therefore be used 
with less hesitation. 

A lutenist accompanying a singer, and asked by him to transpose up 
or down by a small interval, may be better advised to change his tuning 
to a lower or higher actual pitch instead, in order to avoid this loss of 
resonance. But he cannot do this repeatedly in a single concert, because 
of the unsettling effect on the tuning it will not hold its pitch if changed 
too frequently. 

A larger transposition, for example by a fourth or a fifth, is not only 
more worth making, but has the further advantage that violins, etc., 
can pass, e.g. from D major to A major without much affecting the 
number of open string giving support by sympathetic resonance. 





It is a surprising fact of nature, but true, that acoustically just intervals 
will not fit with one another. 

If we pass a just fifth up from C, we reach G; from there a just 
fourth will take us down to D. 

If we pass a just major sixth up from C, we reach A; and from here 
a just fifth will again take us down to D, 

But not quite to the same D. It is a little flatter. 

Again, twelve just fifths add up to the same as seven octaves 
almost. They make, in fact, a little more. 

Such discrepancies make 'perfect* intonation an impossibility. Good 
performers keep up a continual flexibility in their intonation, easing 
each interval into as just a relationship as they can, and paying no 
regard to where the sum-total of these instinctive adjustments may be 
leading them. That is why choirs (having no open strings, etc., to limit 
their adjustments) so often end at a different pitch from their starting- 
point. Nature rather than the choir should take the blame for this. 

These flexible adjustments are temperament but temperament un- 
conscious and in motion. Instruments rigidly pre-tuned, i.e. chiefly the 
keyboard instruments, cannot adjust flexibly and in motion. They have 
to be systematically tempered, in advance. D can only be one kind of 
tone's distance above C; twelve fifths have somehow to be condensed 
into seven octaves. 

Temperament is thus a compromise for spreading the unavoidable 
adjustments as favourably as possible. Keyboard temperament is a 
static temperament, which if it flatters some tonalities, can only do so 
at the expense of others. 


Late renaissance and baroque temperaments could flatter the 'near* 
keys because the 'remote' keys were not in common use. 

A favourite was mean-tone temperament, in which the mean is 
struck between the tones C to D as reached by the different routes 
described in 1 above. Just thirds are possible, and some just fifths; 
but where the cycle of fifths breaks (between G sharp and A flat, or 



D sharp and E flat), the accumulated mistuning the 6 woif is too 
much to be endured. 

Hence mean-tone temperament is very good indeed in some tonalities, 
terrible in others. 


Among the comprehensive range of temperaments explored by late 
renaissance and baroque theorists and musicians, some of whose 
modulations pass through the entire cycle of keys, there were various 
modifications of mean-tone temperament. For example, by the slight 
sacrifice of sharpening the thirds to become a very little wide, the 
fifths can be widened too, and the gap in the cycle narrowed until the 
wolf howls much less noticeably. J. Murray Barbour ('Bach and The 
Art of Temperament', Mus. Quart., XXXIII, January 1947, p. 64) has 
shown that this version of mean-tone temperament with sharp thirds 
(and not equal temperament as historians have assumed) is almost 
certainly the temperament for which J. S. Bach showed his practical 
enthusiasm by composing his Forty-Eight Preludes and Fugues and 
much other music in tonalities too remote for normal mean-tone 


Equal temperament, a necessity now because of our unrestricted 
modulations of tonality, is reached by dividing the octave into twelve 
equal semitones regardless of all resulting mistunings. Every note can 
be taken enharmonically: we can reach the same note as G sharp and 
leave it as A flat, etc. No interval is really good, but none is really bad, 
and the howling of the wolf is for ever silenced. 


(a) Experiments made by A. R. McClure (see Galpin Soc. Jour., No. 1, 
London, March, 1948) and others show that mean-tone temperament 
brings very appreciable advantages for music in a sufficiently restricted 
range of tonalities. 

A harpsichordist who could keep an instrument in mean-tone tem- 
perament (no normal tuner at present could do so for him) would do 
early music an interesting and valuable service by using this in solos. 
The difference, however, is not so marked as to be regarded as an 
interpretative necessity. 

In ensemble music, the difference would not be greatly noticed. The 
melodic instruments would still pursue their own fluid compromise in 
the direction of just intonation, as they do now when playing with 
equally tempered keyboards. 



(b) The case is somewhat different with the organ. Because of its 
sustained tone and its ranks of fifths and other mixtures, the 
beats very harshly in equal temperament, and sounds quite astonishingly 
purer, sweeter and more glowing in mean-tone temperament. 

It would be very greatly to the advantage of early organ music if, 
for example, the parts of the Royal Festival Organ intended as baroque 
replicas could be tuned to mean-tone temperament. Even with smaller 
organs the possibility of isolating some stops for this purpose would 
be worth considering. An organist prepared to forego later organ music 
could insist on mean-tone throughout. The 'sharp' variety thought to 
have been used by J. S. Bach (see 3 above) would be another possibility, 
with the important advantage of being more versatile. 



The Voice 


Teachers of singing are legendary for their jealous conservatism. We 
find Stephen de la Madeleine's Theorie du Chant (Paris, 1864) frankly 
copying Tosi's Opinioni of 1723. We read nineteenth century accounts 
of vocal feats which remind us of Bumey and the eighteenth. We hear 
the rare modern spokesmen of Italian tradition, such as Rupert Bruce 
Lockhart in London, talking like a page from Mersenne in 1636, de 
Bacilly in 1668 or Rameau in 1760. 

There is only one fundamental technique of singing: what we now 
call the old Italian technique. But to Wagner, it did not seem either 
particularly old or exclusively Italian. It was simply good singing. The 
technique was in fact Italian; the resulting alliance with German inter- 
pretation lasted until the generation trained pre-1914 began to thin 
out. There are few singers of a generation younger than Ebe Stignani 
whose voice production carries full conviction. This does not only 
apply to Italian roles. It is becoming as hard to cast Wagner as Verdi 
or Bellini. The same basic requirements apply to Monteverdi or Bach. 

Italian technique includes placing the vocal apparatus precisely; 
supporting the voice from the chest rather than directly from the 
diaphragm; using little rather than much breath; projecting vowel- 
sounds from the upper resonances in detachment from the voice; 
blending the so-called registers; preparing the placing for the next note 
before the last is left. 

As we listen to the last generation of Italian singers on their primitive 
gramophone recordings, often made at an age when a faultily produced 
voice would have already disappeared and when even a superbly pro- 
duced one is long past its best, we hear distantly through the mechanical 
barrier a limpid flow of incredible agility and accuracy, not one note 
of which either forced or out of scale. No effort is made at volume; it 
is the placed voice which reaches the back of the hall, the forced voice 
which fails to carry. The line is moulded, the voice is placed, with 
almost superhuman control. 

The interpretation is so intrinsically vocal that it cannot be conceived 
apart from its technique. The smooth, deep current of sound never falters, 
but its surface is a continual ripple of changing colours, modulations 



of volume, flexibility of tempo. The imagination is pure singer's 
imagination. And not only each note, but word is to live. 

The vowels are exploited for their varied colourings, the consonants 
for their variety of articulation and their full declamatory value. 
Caressed in this way fiercely and gently by turns, the words add another 
dimension to the music. 

There has been some, but not yet enough recovery in the technique 
of singing since the Second World War. 


(a) The following descriptions record general impressions made by 
baroque singers on their contemporaries. 

(667) M. Mersenne, Harmonic Universelle, Paris, 1636-7, II, vi, 

*[A solid sostenuto without pitch-wobble; flexibility in passage-work; 
accurate intonation;] sweetness and a certain harmoniousness, on which 
depends the charms which ravish the hearers, for voices which are hard 
do not please, however accurate they may be, and possessed of the other 
qualities I have mentioned, for they have too much sharpness (aigreur), 
and glitter (esclat), which hurts sensitive ears, and which hinders their 
gliding pleasantly enough into their hearers' spirit to win them, and to 
carry them whither so even you desire/ 

(668) Le Cerf de La Vieville, in Histoire de la Musique, Amsterdam, 
1725, II, p. 305: 

'A perfect voice should be sonorous, extensive, sweet, neat, lively, 
flexible. These six qualities which nature assembles but once in a century, 
are usually found bestowed by halves/ 

(669) Charles Burney, citing a German correspondent, Present 
State of Music in Germany, London, 1773, II, p. 108; 

* [Of the German singer SchmeUing:] I never knew a voice so powerful 
and so sweet, at the same time: she could do with it just what she pleased. 
She sings from G to E in altissimo, with the greatest ease and force, and 
both her portamento di voce, and her volubility are, in my opinion, 
unrivalled; but when I heard her, she seemed to like nothing but difficult 
music. She sang at sight, what very good players could not play, at sight, 
on the violin; and nothing was too difficult to her execution, which was 
easy and neat. But, after this, she refined her taste, insomuch that she was 
able to perform the part of Tisbe, in Hasse's opera, which requires sim- 
plicity and expression, more than volubility of throat; and in this she 
perfectly succeeded, as Agricola, the translator of Tosi's,4rte de Canto, and 
our best singing master in Germany, assures me. 

4 [Burney comments, p. Ill:] Her voice was sweetly toned, and she 
sang perfectly well in tune. She has an excellent shake, a good expression, 



and facility of executing and articulating rapid and difficult divisions, that 
is astonishing . . . she was by no means lavish of graces, but those she used, 
were perfecdy suited to the style of the music, and idea of the poet. 

*[p. 188: the dramatic soprano Faustina:] her execution was articulate 
and brilliant. She had a fluent tongue for pronouncing words rapidly and 
distinctly, and a flexible throat for divisions, with so beautiful and quick 
a shake, that she could put it in motion upon short notice, just when she 
would. The passage might be smooth, or by leaps, or consist of iterations 
of the same tone, their execution was equally easy to her, as to any 
instrument whatever/ 

(b) The following quotations have a more direct bearing on 

(670) Marin Mersenne, Harmonie Universelk, Paris, 1636-7, Livre 
Premier de la Voix, p. 40: 

It is a certainty that the windpipe of the larynx (1'anche du larynx), 
that is to say the little tongue (languette) or its opening, contributes more 
directly to the [performance of florid] passages and divisions than the 
other parts, in as much as it has to mark the degrees and the intervals 
which are made in sustaining the passage; which can only occur from the 
different openings of the little tongue (languette), as I have shown in 
speaking of low and high sound. Thence it follows that those whose little 
tongue (languette) aforesaid is the most mobile, are the most suited to 
make passages and divisions, and that those in whom it is too hard and dry 
cannot make them. Now passages and divisions can be made either in the 
gullet (gorge) by means of the windpipe (anche), as I have said, or with 
the lips; but this last manner is distorted, and condemned by those who 
teach good singing. But of all the nations who learn singing, and who 
make passages of die gullet (gorge), the Italians themselves who make, a 
special profession of Music and of declamation (des recits), avow that the 
French make passages the best . . . 

'It is necessary for the muscles and the cartilages to be very responsive 
. . . those who make passages easily have a softer windpipe (anche) . . . 
[p. 44:] As to the Hard and rigid voice, it acquires this vice from the 
hardness of the windpipe (anche)/ 

(671) B. de Bacilly, UArt de Bien Chanter, Paris, 1668, p. 137: 

4 [The Port de Voix is] the transition which is made by a stroke of the 
throat (coup de gosier) from a lower note to a higher and consists of the 
lower note to be sustained; the doubling from the throat, which is made 
on the upper note; and the sustaining of the same note after it has been 
doubled [i.e. subjected to a mordent]/ 

(672) Joachim Quantz, Essay, Berlin, 1752, XII, 52: 

'A singer who [uses his chest voice for runs] is hardly able to perform 
them as rapidly as one who [uses his head voice], although the first will 
always exceH die second in clarity, particularly in a large place/ 



(673) Jean-Philippe Ranieau, Code de Paris, ! 760, 

p. 16: 

'Yes, all the perfections of singing, all its difficulties, depend solely on 
the breath which conies out of the lungs. 

'The larynx, the windpipe and the glottis are not at our disposal, we can- 
not see their different positions, transformations, to each sound we wish 
to give; but we do at least know that they must not be constricted in 
these differences, that they must be left at liberty to follow their natural 
movement, that we are only masters of the breath, and that in conse- 
quence it is for us to govern it so well, that nothing can disturb the 

'Immediately the breath is given with more force than the sound 
demands, the glottis closes, as when the reed of an oboe is pressed too 
much; if this excess of force is further given too precipitately, it stiffens 
the walls of the glottis, and robs it of all its flexibility; on the other hand, 
a torturing (gene), a constraint caused by attention to good grace, to 
gesture, to the taste of the song, to the inflections of the voice themselves 
efforts of which an acquired habit prevents one from being aware 
these arc the true obstacles to beauty of sound, as well as to the flexibility 
of the voice: the sound(?) tient pour lors dupeigne, de la gorge, du canard; 
the voice trembles and produces nothing more pleasant than bleating (le 

[p. 17:] 'We give our attention, then, solely to spinning the sounds 
(filer les sons) by semitones, both ascending and descending, and when the 
habit is a little familiar, the exercise is extended by a semitone on each 
side, then at the end of two, four or eight days, again a semitone, and 
even so up to the impossible: you will be greatly astonished, after two 
months of this exercise at the most, several hours a day, to find the voice 
perhaps extended by two tones at either end . . . 

'When one feels a slight mastery at this exercise, one notices the degree 
of breath during which the sound has its greatest beauty, whether for 
power, or whether for colouring (timbre) ; one returns to it frequently, 
one tries to give this sound at the first impact of the breath (coup de vent), 
without forcing and without constriction; at last time brings the happy 
day when one reaps the success of one's efforts . . . 

'Arrived at this last point of perfection, the rest is mere child's play; 
one tries florid passages (roulades) . . . trills . . . appoggiaturas taken 
ascending; slides are born from there, and these are the sources of all the 
ornaments of singing. 

'The principle of principles is to take pains to take no pains . . . 

'The breath must be so managed during the florid passages as not to 
force them beyond what one feels one can do without forcing oneself, 
diminishing the breath, and consequently the power of the sound, in the 
measure in which one increases the speed; yet to give more breath a few 
days later, to test if one can do so without forcing; then finally to augment 
and diminish it alternately during the same florid passage, to accustom 



oneself to give, so to speak, the shades of the picture, when the expression 
or sometimes even the simple taste of the song, demands it. The same 
applies to trills and appoggiaturas. 

'Roualdes, trills and appoggiaturas are made all in one breath, in the 
same way as the spun sound (son file), so that one does not at once take 
them at one's ease; but in the degree to which the thing becomes familiar, 
the duration of a single breath grows substantially. 

'While the breath flows, one feels the sounds following one another at 
the opening of the glottis; but however little one tortures oneself, the 
glottis suffers for it; closes instead of opening; and what one should feel 
at its opening is felt for the moment at the bottom of the throat, whence 
are born the sounds of the throat (sons de gorge), etc., of which I have 

(674-5) Charles Burney, Present State of Music in Germany, London, 
1773, II, p. 174: 

'[Francesco Bernardi, called Senesino] sung rapid allegros with great fire, 
and marked rapid divisions, from the chest, in an articulate and pleasing 

'[p. 179: Gaetano Orfmi's] shake was perfect and his portamento, 
excellent. In allegros, he articulated divisions, particularly in triplets, 
most admirably, and always from the breast [chest]. In adagios he was so 
perfect a master of every thing which pleases and affects, that he took 
entire possession of the hearts of all that heard him; he was many years in 
the imperial service, and though he lived to an advanced age, he pre- 
served his fine voice to the last [a sure sign of good voice-production. A 
footnote adds:] He died at Vienna, about the year, 1750. 

'[p. 241 : of "one of the BagBoni", her compass] B on the fifth space in 
the base, to D in alt, fully, steady and equal; her shake is good, and her 
Portamento admirably free from the nose, mouth, or throat [i.e. ? not 
produced from those organs]. There was such a roundness and dignity in 
all the tones, that every thing she did became interesting; a few plain notes 
from her, were more acceptable to the audience than a whole elaborate 
air from any one else/ 


(676) M. Mersenne, Harmonie Universelle, Paris, 1636-7, II, vi, 355: 

'Now when voice-production has been taught (apres que Ton a 
enseigne a former le ton), and the adjustment of the voice to all sorts of 
sounds, familiarity with the making of cadences is imparted, which corre- 
spond to the trills and mordents (aux tremblemens et aux martelemens) 
made on the keyboard of the organ and the spinet, and on the finger- 
board of the lute and other stringed instruments. These [trilled] cadences are 
the most difficult part of all that has to be done in singing, because it is 
necessary simply to beat the air in tke throat (battre Fair de la gorge), 



which must make a series of repercussions (tremblemeos) without the 
help of the tongue. But they are as much more pleasing as they are more 
difficult, for if the other progressions are the colours and the shadings, the 
cadences can be called the rays and the light. Those who have not the 
constitution of the throat for the said cadences and florid passages, make 
use of movements of the tongue, which are not so pleasing, especially 
when they are made with the tip; for as to those which are made with the 
middle, they are necessary for certain passages which cannot be performed 
accurately enough without the help and repercussion (tremblement) of 
the middle of the tongue, because of the occurrence of the vowels, 
which must be pronounced, and rendered intelligible to the hearers. 

"As for ornaments from the lips, they are not agreeable, nor permitted, 
any more than those which seem to be drawn from the stomach.' 

We may compare with the above early baroque passage the two 
modern ones next following. 

(677) Franklyn Kelsey, Grove's Dictionary, 5th ed, London, 1954, 
Vol. IX, s.v. 'Voice-training 1 , p. 65: 

It is a fact known to all competent teachers that the trill cannot be 
produced by orthodox vocalization, for the tensioning muscles of the 
glottis simply will not oscillate rapidly enough to cause it. An attempt to 
trill by orthodox vocalization always results in what the Germans call a 
Bockstriller [and the French chevrotter]* sound like a goat-beat. The trill 
is in fact caused by a controlled oscillation of air-pressure in the windpipe, 
the trachea acting like an accordion, stretching and shortening itself with 
extreme rapidity and so causing an oscillation of the entire larynx/ 

(678) An 'old Italian singing-teacher in New York', quoted in a 
letter to Musical Opinion, July, 1946, and cited by Kelsey (loc. a'/., p. 

'You modern singers always seem to want to make your notes in your 
mouths. We old singers always made our notes in our throats, long before 
they came into our mouths !' 

The rapidity of the trill is, of course, only relative. A very rapid 
vocal trill is necessarily tight and narrow, and is always to be avoided. 

(679) Charles Burney, Present State of Music in Germany, London, 
1773, II, p. 137: 

'[Anfani's] shake is a little too close, otherwise I should venture to 
pronounce him a perfect tenor singer.' 

The following suggests another view of the trill: namely that it is 
not produced by an 'unorthodox' technique, but by an intensive 
cultivation of normal agility. The authority cited is the great teacher 
Mancini; but Mancini may not have realised that agility carried to this 



point does change its technical basis in the way described by Kelsey at 
(677) above. That would, indeed, account for the difficulty found with 
the trill by 'many who can execute passages'. 

(680) Charles Burney, Present State of Music in Germany, London, 
1773, I, p. 334: 

'For the shake he [Signor Mancini] thinks it ruined ninetynine times 
out of hundred by too much impatience and precipitation, both in the 
master and scholar; and many who can execute passages, which require 
the same motion of the larynx as the shake, have notwithstanding never 
acquired one. There is no accounting for this, but from the neglect of the 
master to study nature, and avail himself of those passages which, by 
continuity, would become real shakes.' 


A confusion of terms exists among singers: vibrato is used by some 
for a desirable oscillation of intensity (strictly a tremolo) which corre- 
sponds, in singing, to the instrumental vibrato; by others for an un- 
desirable oscillation of pitch (strictly a vibrato). 

An uncontrolled pitch-wobble is a distressing disturbance of tech- 
nique which can arise from faulty voice-production, and which once it 
has arisen is appallingly hard to cure. It is condemned by all teachers. 
Not all teachers, however, realise that the true vocal 'vibrato' is not 
only an intensity fluctuation but also (though less conspicuously) a 
pitch fluctuation. The 'wobble' sounds weak, vacillating and out of 
control. The true vocal Vibrato' sounds wonderfully alive, decisive and 
under control. It is an essential element of voice-production, as much 
so in early music as elsewhere. See also Ch. XIX above. 


The traditional Italian portamento is practised as if it were on definite 
notes: chromatic if the interval bridged is less than about a fifth; 
diatonic if it is more. It is (except as a special coloratura effect) sung 
by running the notes together, but remains under control, and is neither 
a wild scoop nor a premature collapse. The vocal apparatus is aimed 
towards the higher pitch in ascending, and held well up even in de- 
scending. The portamento itself ordinarily comes rather quickly, and 
should always sound (like the true vocal 'vibrato') as steady as a rock. 

Caccini at (660) in Ch. LI, 5, above, refers to a slide from the third 
below, which in singing might mean the more or less sotto voce porta- 
mento from below with which so many notes are habitually attacked. 




The references to the chest voice at (672) and (674) above are exceed- 
ingly important. There can be no bel canto without an excellent chest 
voice to colour the bottom register and to blend upwards into the 
medium and head registers above. 

(680a) Claudio Monteverdi, letter to the Duke of Mantua, 9 June 

*A fine voice, strong and long [sustained]; and singing in the chest, he 
reached all places [in the voice] very well/ 

(680b) Joachim Quantz, Autobiography (1754) in Marpurg's His- 

torisch-kritlsche Beytrage, I, Berlin, 1755, pp. 231-232: 

'[Paita's tenor voice] would not have been by nature so fine and even, 
if he himself, through art, had not known how to join the chest voice with 
the head voice. 

[p. 235 : 'Carestini] had a great dexterity in [ornamental] passages, which 
he, according to the good school of Bernacchi, like Farinello, produced 
with the chest/ 

It is skilled chesting which gives easy power, brightness and clarity 
to sopranos and mezzo-sopranos high up, transparent strength, sureness 
and colour low down. The colour is strong and not now in fashion, but 
will be admired again if we can bring it back (Stignani retained a good 
deal of it, and indeed it is particularly needed for the mezzo voice). 
Men have usually enough chest voice, but have a problem in blending 
it into their high notes (it is this which gives the Heldentenor his true 
ring, and which distinguishes a genuine bass-baritone). 

Voice teachers have a valuable part to play in developing this difficult 
but necessary reform. Strong chesting has its dangers, if the voice is 
forced; but not if it is placed and focused accurately, well supported 
and brought sufficiently forward. It will then sound almost quiet close 
by, but very open, and will carry through an orchestra to the back of 
the theatre at sixty years of age. 


Castrato voices are now neither extant nor in prospect. The best 
substitute when the part does not lie too high is a good counter-tenor, 
for which see 8 below. For high castrato parts, a dramatic soprano is 



the best substitute. A boy soprano has neither the power nor the 
musical sophistication to be a suitable alternative. 

Since counter-tenors are also in short supply, the lower castrato parts 
may also have to be taken by women. A contralto is always a possibility 
here, provided her technique is light. 

(681) Jean- Jacques Rousseau, Dictionnaire de musique, Paris, 1768, 
s.v. 'Castrato' : 

'The advantage of the voice is off-set in the Castrato by many other 
disadvantages. These men who sing so well, but without warmth, and 
without passion, are, in the theatre, the most unmannerly actors in the 
world; they lose their voice in very good time and acquire a disgusting 
fatness. They speak and pronounce worse than true men, and there are 
even letters such as r, which they cannot pronounce at all.' 

(682) Charles Burney, Present State of Music in Germany, London, 
1773, II, p. 179: 

'Domenico had one of the finest soprano voices ... it was so clear and 
penetrating, as to make its way through all obstructions, and with all 
this great force, was sweet, and well toned; however he neither sung nor 
acted with much spirit.' 

The last known castrato singer, Alessandro Moreschi, died in 1922, 
after making a gramophone recording which Desmond Shawe-Taylor 
has heard and described, not in very enthusiastic terms. The tradition 
had no doubt gone far down by then. 


(a) A male alto is a bass or baritone singing high parts in a falsetto 
voice, apparently produced by using only the edges of the vocal cords. 
He is by tradition a valuable member of English cathedral choirs. 

(b) The counter-tenor is a natural voice often of extremely high 
tessitura, light and more or less unchested yet masculine in quality. 
The range is from g to d 11 or even & . The consummate artistry of Alfred 
Deller has revived it in this generation, and others have followed: it is 
a very necessary voice in much music by Purcell and others. 

(c) A normal tenor may sometimes go extremely high, as in the 
magnificent voice and musicianship of Hugues Cuenod. 


Most early texts are casual, incomplete and inaccurate in their under- 
laying of the words; it is for the performer to decide which syllable 
should go to which note or notes. 

One curious but pleasing difference from modern methods has been 
thoroughly established by E. H. Fellowes (English Madrigal Composers, 



Oxford, 1921, p. 112) for the Elizabethan and early Jacobean school; 
it was generally characteristic of the late renaissance and early 
periods, and may be seen as follows. 

Ex. 225. Elizabethan underlaying, cited E. H. Fellowes, for. ri/., 
(a) Gibbons, Christ Church, Oxford, MS. 21; (b) Weelkes Ayeres, 

London, 1608, No. 2; (c) Tomkins Songs, London, 1622: 


Rc-joicrin Him, it. -- Hun lor lit- nit. 

hath re-stored to us and by His ris - ing , to life a -gain. 



Jock-ie thine hornpipe* 1 ; dull. 

JfcJ J ^jjf I ff *J 4J 
iff ~ .. -^>rr t. L<f l . l -.'Nj,jra* 

our life a-\\uy doth post, a - way 


(683) Thomas Morley, Plaine and Easie Introduction, London, 1597, 
p. 179: 

"Our church men . . . ought to studie howe to vowell and sing clean, 
expressing their words with devotion and passion.' 

(684) Charles Butler, Principles of Musik, London, 1636, p. 91: 

'[Singers should] sing as plainly as they would speak: pronouncing 
every Syllable and letter (specially the Vowels) distinctly and treatably.' 

(685) Le Cerf de La Vieville, in Histoire de la Musique, Amsterdam, 
1725, II, p. 25: 

'The Italian singers pronounce badly . . . because . . . they close all 
their teeth and do not open the mouth enough: except in the florid 
passages, where they hold it open for quarter of an hour at a time, without 
moving the tongue, or the lips . . . it is only in France that they take the 
trouble to open the mouth as they should in singing/ 

The above view was biased against the Italians, being expressed in 
furtherance of the national controversy opened up by Raguenet's 
Comparison of the French and Italian Music in 1702; but that French 
singing was particularly declamatory is confirmed from other sources, 
as also that it became in some respects less so at the end of the baroque 
period (Quantz would be less up to date than Rousseau here): 



(686) Joachim Quantz, Essay, Berlin, 1752, XVIII, 76: 

'The Italian manner of singing is refined and full of art, it moves us and 
at the same time excites our admiration, it has the spirit of music, it is 
pleasant, charming, expressive, rich in taste and feeling, and it carries the 
hearer agreeably from one passion to another. The French manner of 
singing is more plain than full of art, more speaking than singing; the 
expression of the passions and the voice is more strange than natural/ 

(687) Jean- Jacques Rousseau, Lettre sur la Musique Frangoise [Paris], 
1753, p. 53: 

'The old recitative was performed by the actors of that time quite 
otherwise than we do it today; it was more lively and less dragging; they 
sang it less, and they declaimed it excellently/ 

These variations of period and nationality do not diminish our 
impression that a large proportion of baroque delivery was more 
highly dramatised than we tend to make it. The florid passages were 
slipped in lightly and easily; the cantabile never wavered; but the 
enunciation was incisive, especially on the consonants, and full of both 
musical and v